Foreign Exchange Interventions as an (Un)conventional Monetary Policy Tool: Mid-Term Evaluation

Transcription

1 11th International Conference Challenges of Europe: Growth, Competitiveness and Inequality Foreign Exchange Interventions as an (Un)conventional Monetary Policy Tool: Mid-Term Evaluation Lubomír Lízal, Ph.D. Hvar, May 7, 15

2 Outline 1. Situation of the Czech economy in 13. Passive monetary policy vs. (un)conventional tools 3. FX commitment: An additional instrument for easing the monetary conditions 4. FX commitment: Reflection and mid-term evaluation 5. Summary and conclusions

3 Outline 1. Situation of the Czech economy in 13. Passive monetary policy vs. (un)conventional tools 3. FX commitment: An additional instrument for easing the monetary conditions 4. FX commitment: Reflection and mid-term evaluation 5. Summary and conclusions 3

4 Situation of the Czech economy in 13: Inflation actual vs. targets (y/y; %) Inflation was decreasing and was at the lower boundary of the tolerance band around the CNB s target (below 1%) at the end of 13. In spite of VAT and other tax increases. MP inflation was around zero and also showed a decreasing tendency. 4

5 Situation of the Czech economy in 13: Core inflation tradables and non-tradables (y/y; %) 15 1 Adjusted inflation (except fuels) Prices of non-tradables (except administered prices) Korigovaná inflace bez pohonných hmot (PH) Neobchodovatelné-ostatní Obchodovatelné-ostatní Prices of other tradables bez (except PH food and fuels) /94 1/97 1/ 1/3 1/6 1/9 1/1 Core inflation negative since 9. For the first time in modern history the growth of prices of nontradable goods (and services) stopped. This is very unusual, especially for a converging economy. Partly, this was caused by a decline in prices of communication services. But primarily, it reflected weak demand and slow wage growth. 5

6 Situation of the Czech economy in 13: Structure of GDP growth (y/y; p.p.) I/9 I/1 I/11 I/1 I/13 Household consumption Gross fixed capital formation Change in inventories Net exports Government consumption NPISH expenditure Czech economy back in recession since the beginning of 1. Net exports had a positive contribution only in Q13. Fixed investments were negative, change in inventories fluctuated. The drop in consumption has stopped. 6

7 Situation in November 13 Inflation Report IV/13: Interest rate forecast Consistent with the forecast was a significant decline in market interest rates well below zero. (3M PRIBOR in %) IV/11 I/1 II III IV I/13 II III IV I/14 II III IV I/15 II 9% 7% 5% 3% confidence interval Consistent with the forecast was a significant decline in market interest rates well below zero, followed by a rise in rates above the current levels only at the end of 14. But W repo rate lowered to.5% on November 1, 1! Given the zero lower bound (ZLB) on interest rates, this pointed to a significant need to ease monetary policy using other instruments. 7

9 What would be the implications of passive MP? The key question in this situation was, how would the economy evolve if no policy action was taken. One could not reasonably expect that the exchange rate would not move with rates staying at zero and no interventions, and that all other macro-economic variables would evolve in line with the baseline scenario, too. No action from the CNB, once all previously communicated triggers have been crossed, would erode credibility of verbal interventions, and passivity of monetary policy would most probably become fully expected. A scenario with passive monetary policy was thus simulated, actually in two versions differing in the severity of their impact. This is a what-if, counterfactual scenario, i.e. something you will not see in reality, but should keep it in mind when assessing policy. 9

10 Consequences of passive MP 4 Headline inflation (y/y in %) 7 CZK/EUR I/1 III I/13 III I/14 III I/15 III 4 I/1 III I/13 III I/14 III I/15 III 1,4 1, 1,,8,6,4,, -, -,4 -,6 3M PRIBOR (in %) I/1 III I/13 III I/14 III I/15 III GDP (y/y in %) Baseline scenario Passive MP Passive MP +ER shock I/1 III I/13 III I/14 III I/15 III Inflation would get negative for at least two to three quarters, the CZK s exchange rate would appreciate, and the recovery would get delayed. Downward risks: in the ST horizon an exchange rate appreciation shock; in the LT lost gravity of the inflation target for expectations. (Equilibrium: In the model the inflation always returns back to the target, but in reality?) Moreover: Cook & Devereux (13) show that the optimal reaction of foreign policy rates to home rates at ZLB is an increase otherwise the home currency starts to appreciate uncontrollably. But ECB decreased its rates -> further appreciation pressure on CZK. 1

11 (Un)conventional monetary policy tools Central banks faced the limits of the standard monetary policy instrument due to long-lasting stagnation. Turned their attention to various unconventional MP measures negative interest rates, liquidity provision (QE), FX interventions. Most economies have been hit by liquidity crisis and subsequent credit crunch, but Czech financial system has abundance of liquidity. Ratio of deposits to loans above EU average, total deposits exceed total loans. Excess liquidity absorbed by the CNB using repo tenders. Liquidity ratios over time (%; QA = quick assets) Ratio of deposits to loans granted in selected EU countries (%; end of 11; deposits/loans to residents) QA / assets 3 Client deposits / loans (rhs) 11 QA / assets (trend) 15 Client deposits / loans (trend, rhs) 1 3/1 1/1 9/11 6/1 3/13 1/ CZSKPLROBGHU SI LT EELV BEATFR IT UKSEDK EAEU Source: CNB Inflation Reports; ECB 11

12 (Un)conventional monetary policy tools In 8, the CNB introduced liquidity-providing repo operations, but they were used very rarely. Experience with negative deposit interest rates is very limited, the consequences in general very unclear. Also potential legal complications with negative rate (penalty interest bound in multiplicative manner to the discount rate). Therefore, FX interventions selected by the Board as the most appropriate MP tool when interest rates hit zero. W repo rate lowered to.5 % on November 1, 1. Open market operations and currency in circulation (CZK billions) Monetary policy rates since the start of the financial turbulence (%) Liquidity-withdrawing repo operations Currency in circulation Deposit facility Liquidity-providing repo operations (rhs) 3/7 3/8 3/9 3/1 3/11 3/1 3/13 3/14 Source: CNB Euro area US CZ

13 Compatibility of FX interventions and inflation targeting Inflation targeting used by many central banks as the means of achieving and maintaining price stability (CNB since 1998). Under IT, the MP tool is the interest rate exchange rate is an endogenous variable affecting inflation expectations. Are FX interventions compatible with IT? Leaving aside technical and procedural compatibilities, does influencing FX compromise the credibility and goals of the CB? MP implicitly reacts to exchange rate shocks under standard conditions: An unexpected appreciation shock tightens the monetary conditions; Interest rate is lowered to easy the conditions again. 13

14 Compatibility of FX interventions and inflation targeting According to the IMF s classification, independent floating dominated among IT countries before the crisis (with 19 countries as of April 8), but managed floating coexisted with this regime, too (1 countries). Many of officially independent floaters do occasionally intervene, with interventions more common during the recent crisis: Brazil, Chile, Indonesia, Israel, Mexico, New Zealand, Poland, South Korea, Switzerland. The literature advocates freely floating exchange rate for IT regime but doesn t deal with situation when interest rates hit the zero lower bound (ZLB). FX interventions with (almost) zero interest rates: Switzerland & Japan (both independent floaters) 14

15 Exchange rate pass-through to consumer price inflation Existing empirical evidence suggests that FX interventions have the ability to influence the exchange rate, but not necessarily in the long term. Fatum & Pedersen (9): Sterilized FX interventions in Denmark have a significant influence on exchange rate returns when the direction of intervention is consistent with the MP stance. This is the case when an FX shock is NOT accommodated but rather fueled by MP inactivity. When the interest rate hits zero and can no longer be used or is insufficient to influence inflation expectations and achieve price stability, FX interventions become a viable option. Svensson s (1, 3) way to escape the ZLB 15

16 Exchange rate pass-through to consumer price inflation But how do exchange rate changes affect prices in the Czech Republic? Is the existing evidence of exchange rate pass-through relevant for FX interventions? The estimates of the transmission of an exchange rate shock to Czech inflation lie between % to 8%. In the standard inflation-targeting regime, a shock transmits: directly through import prices; indirectly through economic activity: real volumes of imports and exports -> wages -> employment; moderated by reaction of interest rates: correction of nominal exchange rate & through real interest rates influences investment and private consumption Exchange rate pass-through to Czech inflation based on a literature review 1 Pass-through (%) Note: The figure shows the reaction of consumer prices to an exchange rate shock of 1%. It summarizes 43 ERPT estimates for the Czech Republic collected from papers and articles published in 1 1. Dark blue points represent time-invariant estimates. Light blue points are medians of time-varying estimates. Results based on the CNB RPN (forthcoming) are in dark red (VAR, VECM) or red (BVAR median and TVP-VAR for 8Q1, 9Q1 and 1Q1). Source: CNB Last observation (year) 16

17 Influence of the zero lower bound on pass-through As rates approach the ZLB, the transmission of shocks to the economy may change. Bayesian fan charts (Franta et al. 13): Forecasts can be conditioned on shocks that lead to a non-negative nominal interest rate (Fig. 1). Forecasts can be conditioned on the interest rate itself (regardless of shocks influence on other vars). ZLB can be ignored (Fig. ): ex-post observed MP relevant inflation on the edge of the centered 95% of the distribution forecast; over-optimistic growth outlook. Fig. 1: Shocks filtered (ZLB active) Fig. : ZLB ignored Source: Franta et al. (13) 17

18 Influence of the zero lower bound on pass-through In general, fixing one input or binding constraint leads to higher responsiveness (volatility) of the remaining variables. That is, when MP reacts transparently to an exchange rate shock, the pass-through to inflation is relatively small. But as the CB approaches the ZLB and interest rates can t counteract the shock, the transmission increases sizably depending on how long economic agents expect MP to operate in the ZLB regime: Direct channel through import prices is stronger. Indirectly through real interest rates and expectations about their future development: longer-term fixed nominal interest rates and increasing inflation push real interest rates down. Analogously, fiscal multipliers are higher in a ZLB regime (Gerchert & Will, 1). 18

19 Svensson s foolproof way vs. CNB s FX commitment Svensson (1, 3): Announcing a price-level target path corresponding to a positive long-run inflation target, a devaluation of the currency and a temporary exchange rate peg, which is later abandoned in favor of standard inflation targeting once the price-level target path has been reached. CNB s FX commitment: No explicit price-level target but inflation expected to overshoot the target = implicit temporary price-level target, equivalent from ex ante perspective; Exchange rate weakening as a one-off action, no exchange rate peg; Announcement that a future change of the exchange rate to a weaker level was possible in case of strong anti-inflationary pressures. 19

20 Outline 1. Situation of the Czech economy in 13. Passive monetary policy vs. (un)conventional tools 3. FX commitment: An additional instrument for easing the monetary conditions 4. FX commitment: Reflection and mid-term evaluation 5. Summary and conclusions

21 What did we expect: alternative scenario vs. passive MP 4 Headline inflation (y/y in %) 4 MP-relevant inflation (y/y in %) I/1 III I/13 III I/14 III I/15 III -1 I/1 III I/13 III I/14 III I/15 III 1,4 1, 1,,8,6,4 3M PRIBOR (in %) GDP (y/y in %) Baseline scenario Passive MP Passive MP +ER shock Alternative scenario,, -, -,4 -,6 I/1 III I/13 III I/14 III I/15 III I/1 III I/13 III I/14 III I/15 III The relevant policy comparison is between the alternative scenario with exchange rate instrument and the passive MP scenario(s) here the difference is clearly favourable for all major variables (plus uncertainty as regards the return to the target with passive MP). 1

22 The Bank Board s decision of November 13 The Board decided to start using the exchange rate as an additional instrument for easing the monetary conditions, stating that: The CNB will intervene on the FX market to weaken the koruna so that the exchange rate is close to CZK 7/EUR. The exchange rate level was chosen to avoid deflation or long-term undershooting of the inflation target and to speed up the return to the situation in which the CNB will be able to use its standard instrument, i.e. interest rates. The exchange rate commitment is one-sided. This means that the CNB will prevent excessive appreciation of the koruna exchange rate below CZK 7/EUR. On the weaker side of the CZK 7/EUR level, the CNB is allowing the exchange rate to move according to supply and demand on the FX market.

23 Interest rate and exchange rate Development of W REPO rate (in %) and CZK/EUR exchange rate CZK/EUR W Repo (right axis) Monetary policy easing & exchange rate depreciation go hand in hand. Impossible trinity: cannot have independent monetary policy, stable exchange rate, and free movement of capital at the same time! 3

24 CZK/EUR rate and the CNB commitment /1 1/11 1/1 1/13 1/14 1/15 CZK/EUR 7 CZK/EUR level After the CNB s policy announcement, koruna reached 7 CZK/EUR quickly, and has been moving at somewhat weaker levels since then. The exchange rate volatility decreased significantly (both the actual one, and implied by futures prices), except for the most recent period. 4

25 CNB s FX operations and the exchange rate Actual interventions were quite massive, but took place only for a few days after the policy decision of the CNB. 5

27 Monetary conditions in the Czech Republic Source: CNB IR II/15 Basic RMCI version (positive values refer to easy monetary conditions and negative values to tight monetary conditions; source: CNB, CNB calculation) ,4 I/4 I/5 I/6 I/7 I/8 I/9 I/1 I/11 I/1 I/13 I/14 I/15 GDP decrease Interest rate component Basic RMCI Real monetary conditions index = 3M PRIBOR market rate adjusted for inflation expectations + exchange rate component. Describes the aggregate monetary policy stance. The monetary conditions have been very accommodative since the end of 13, albeit rather less so recently as a result of a fall in inflation expectations.,3,,1, -,1 -, -,3 Exchange rate component CZK/EUR 7

28 Comparison: Real effective exchange rate (7=1%) Czech koruna Swiss franc 7 average = US liquidity crisis Swiss FX commitment Czech FX commitment 6 Fall of Lehman Brothers 5 1/95 1/97 1/99 1/1 1/3 1/5 1/7 1/9 1/11 1/13 1/15 Source: BIS SNB trying to tame the strong appreciation of Swiss franc after the 8 crisis. Real effective exchange rate of CZK appreciated by 1% during the crisis period, which is equivalent to permanent tightening of monetary policy. Counteracted by interest rate decreases. Verbal interventions only have limited effects and work only for limited time. Still, different reason for intervention regime than in the Czech Republic. 8

29 Real economic convergence in the Czech Republic GDP per capita in PPP (EA-18 = 1) Development CZ stagnation AT growth DE growth PT divergence HU convergence PL convergence SI divergence SK convergence Since the onset of the crisis, long-term real economic convergence stopped and had not yet renewed. Economic convergence was the main driving force behind equilibrium real appreciation of Czech koruna (average annual rate.4% in 4-13). CNB assumes that once the impacts of the crisis have faded away, Czech koruna will renew its real appreciation at a rate of 1.5% a year. 9

30 Situation of the Czech economy in 13: Trade balance surplus and investment (CZK bn.) cross-border concept national concept investment (right axis; inverted) About 6-8% of investment is imported. Yearly nominal investments between 8 and 13 decreased by 17 bn. CZK. Net exports therefore had to improve by about bn. CZK (= 17*.6-17*.85) solely as a statistical consequence. Total trade balance surplus in 13 was 17 bn. CZK. Perceived good results of foreign trade caused appreciation pressures. Already in 1H 14, trade balance surplus +7 bn. CZK and investment +3 bn. CZK compared to 1H 13 -> pattern broken! 3

31 Development since November 13 Annual percentage changes Available on Available on Gross domestic product (s.a.) II/ I/ Consumer price index 9/ /15.5 General unemployment rate (in %, s.a.) 9/ / Average nominal wage II/13 1. IV/14.3 Number of vacancies 9/13 39,4 4/15 83,7 Gross operating surplus of nonfinancial corporations II/ Insufficient demand as a limit of production in industry (in %) IV/ / Composite confidence indicator (index) 1/ / Almost all key macroeconomic indicators visibly changed in positive direction i.e. in accordance with the CNB s predictions. Inflation is lower compared to November 13, but monetary policy-relevant inflation (without primary impact of indirect taxes changes) was.5% in April 15, which is higher than on the beginning of

32 Forecasts of inflation, interest rates, and GDP: Then (IR IV/13) and now (IR II/15) Headline inflation forecast Headline inflation will be close to zero in 15 and rise to the target next year (year on year in %) Monetary policy-relevant inflation forecast Monetary policy-relevant inflation will fluctuate around zero this year and then start to rise, converging to the target at the end of the monetary policy horizon (year on year in %) % 7% 5% 3% confidence interval IR IV/13 Inflation target Monetary policy horizon Inflation target Monetary policy horizon II/13 III IV I/14 II III IV I/15 II III IV I/16 II III IV - II/13 III IV I/14 II III IV I/15 II III IV I/16 II III IV 9% 7% 5% 3% confidence interval Interest rate forecast The forecast expects market interest rates to be flat at their current very low level until the end of 16, i.e. over the entire forecast horizon (3M PRIBOR in %) GDP growth forecast GDP growth will gradually accelerate (annual percentage changes; seasonally adjusted) % 7% 5% 3% confidence interval IR IV/ II/13 III IV I/14 II III IV I/15 II III IV I/16 II III IV -4 II/13 III IV I/14 II III IV I/15 II III IV I/16 II III IV 9% 7% 5% 3% confidence interval Does it mean that the exchange rate tool doesn t influence prices? 3

33 GDP growth decomposition 1,1 Monetary policy + sentiment (+1 pp.) Fiscal impuls (+1. pp.) Foreign demand (+.5 pp.), Growth in 14 (+.%) -,7 Growth in 13 (-.7%) The Czech Republic among the 1 worst-performing countries of the EU up to 13: GDP -1.3% in Q 13. In 14: GDP +%, and the Czech Republic moved among the top 1 EU members. The change in dynamics of the Czech GDP reached.7 pp. in 14 (from -.7% in 13 to.% in 14). Foreign demand + fiscal impulse explain 1.7 pp., the remaining 1 pp. can be attributed to the monetary policy action combined with improved sentiment of businesses and households. 33

34 February 15 sensitivity scenario: No intervention regime Czech GDP (y/y in %) I/14 II/14 III/14 IV/14 Actual GDP Nov 13 CNB forecast Hypothetical GDP had the koruna not been weakened Note: Actual GDP and hypothetical GDP for 14 Q4 are based on the CZSO s flash estimate of GDP. These figures may be subject to revision. Had the monetary conditions not been eased in November 13 using the exchange rate, the economy would have grown by around.5% in

35 February 15 sensitivity scenario: No intervention regime Czech inflation (y/y in %) I/14 II/14 III/14 IV/14 Actual inflation Nov 13 CNB forecast Hypothetical inflation had the koruna not been weakened and inflation would now be running at about -%. 35

36 February 15 sensitivity scenario: No intervention regime Czech household consumption (y/y in %) I/14 II/14 III/14 IV/14 Actual household consumption Nov 13 CNB forecast Hypothetical consumption had the koruna not been weakened The actual impact on domestic household consumption near the predicted levels reveals that the weakening of koruna was successful in helping domestic demand. 36

37 Euro area consensus forecast in November 13 and February 15 (differences in p.p.) Effective GDP in the euro area (annual % change) (differences in p.p.) Effective CPI in the euro area (y-o-y in %) Differences CF February 15 CF October 13 I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV (differences in p.p.) Effective PPI in the euro area (y-o-y in %) (differences in p.p.) 3M EURIBOR (in %) I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV. (differences in %) EUR/USD exchange rate (USD/EUR) (differences in %) Price of Brent crude oil (USD/barrel) I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV 55 Strong disinflationary pressures imported from the euro area. 37

39 Concluding remarks Existing estimates of pass-through are not applicable in a situation of zero rates. Pass-through at the ZLB can be several times larger than past estimates. FX interventions do not reduce the transparency and credibility of inflation targeting. Targeted inflation remains unchanged and is still publicly known in advance. The prediction of the market interest rate path, despite reflecting the zero level, remains publicly announced, too. Exchange rate is still not the target, but rather a new tool for achieving the target. FX interventions used only when the standard tool is no longer available. 39

40 Concluding remarks Power of the CB when intervening against its own currency is not limited by the size of the reserves. Depreciation helps net exports, which is a good sideeffect stimulating the economy. The use of FX interventions to ease the monetary conditions in an open economy is a rational choice at ZLB. FX commitment serves well even if the economic development changes: Efficient tool of monetary easing. Anchor to inflation expectations in case the surrounding development would lead to unwanted monetary tightening. 4

41 The latest CNB s monetary policy decision On May 7, the Bank Board decided to continue using the exchange rate as an additional instrument for easing the monetary conditions and confirmed the CNB s commitment to intervene on the foreign exchange market if needed to weaken the koruna so that the exchange rate of the koruna is kept close to CZK 7 to the euro. In line with this, the Czech National Bank still stands ready to intervene automatically, i.e. without the need for an additional decision of the Bank Board, and without any time or volume limits. The asymmetric nature of this exchange rate commitment, i.e. the willingness only to intervene against appreciation of the koruna below the announced level, is unchanged. The subsequent return to conventional monetary policy will not imply appreciation of the exchange rate to the level recorded before the CNB started intervening, as the weaker exchange rate of the koruna is in the meantime passing through to domestic prices and other nominal variables. 41

42 The latest CNB s monetary policy decision The Bank Board assessed the risks to the new forecast at the monetary policy horizon as being anti-inflationary. Of the three factors which the Bank Board identified at its previous meetings as being crucial in preventing adverse second-round effects of the oil price decrease, domestic wages and the korunaeuro exchange rate have so far been moving in the anti-inflationary direction. By contrast, the optimism about the overall expected benefits of the measures adopted by the European Central Bank for developments in the euro area and in the Czech Republic is growing. In this situation, the Bank Board stated again that the Czech National Bank would not discontinue the use of the exchange rate before the second half of 16. In addition, the Bank Board emphasized again that it was ready to move the level of the exchange rate commitment if needed. 4

43 Thank you for your attention Lubomír Lízal, Ph.D. member of the CNB Board 43

44 Additional slides 44

45 Sensitivity scenario: Higher Brent price Low oil price establishes a positive supply shock for the Czech economy in spite of its anti-inflationary impact. How large a shock? Czech Republic is a net importer of oil products. In 1 and 13, when the average price per one barrel of crude oil moved above USD 1, external trade deficit in oil products was between CZK 1 and 13 bn. If oil keeps to be sold for about half per barrel over a longer period, we can expect a similar drop in external trade deficit by approx. CZK 6 bn. (CNB forecast expects the impact of approx. CZK 5 bn.). This translates to about a one-third increase in trade surplus (CZK 17 bn. in 13, CZK 157 bn. in 14). Almost the same stimulus for Czech economy as the whole state budget deficit (CZK 78 bn. in 14). 45

46 Oil price: Expectations in 13 vs. reality (differences in %) Price of Brent crude oil (USD/barrel) Differences CF February 15 CF October 13 I/13 II III IV I/14 II III IV I/15 II III IV I/16 II III IV Oil price expected to be around 1 USD/barrel but dropped to 6 USD/barrel by the beginning of 15 instead. In order to analyze the impact on Czech economy, CNB constructed a sensitivity scenario with oil price fixed at 1 USD/barrel. 46

47 Sensitivity scenario of the February forecast: Higher Brent price CPI Inflation (y/y, in %) 4 GDP Growth (y/y, in %) I/13 I/14 I/15 I/16 I/13 I/14 I/15 I/16 Household Cons. Growth (y/y, in %) Nominal Wage (y/y, in %) I/13 I/14 I/15 I/16 I/13 I/14 I/15 I/16 Oil price - baseline Oil price - 1 USD/bl. With oil price around 1 USD/barrel, inflation would be on average 1 pp. higher in 15, but the GDP and household consumption growth would be significantly lower. 47