Administered Prices and Inflation Targeting in Thailand Kanin Peerawattanachart

Transcription

1 Administered Prices and Targeting in Thailand Kanin Peerawattanachart Presentation at Bank of Thailand November 19,

2 Jan-96 Oct-96 Jul-97 Apr-98 Jan-99 Oct-99 Jul-00 Apr-01 Jan-02 Oct-02 Jul-03 Apr-04 Jan-05 Oct-05 Jul-06 Apr-07 Jan-08 Oct-08 Jul-09 Apr-10 Jan-11 Oct-11 Jul-12 Apr-13 Jan-14 Oct-14 Thailand has been fairly low after the adoption of inflation targeting in May2000 around the World (in percentage) World Advanced economies Developing economies ASEAN Thailand Source: World Economic Outlook (2014) Headline Core Source: CEIC (2014) 2

3 BoT shifted from core inflation targeting at 0.5-3% to headline inflation targeting at 2.5% ± 1.5% Switching of Targeting Regime Reasons Core Headline Easier communication: more attentive to general public Band Point Better anchoring: one target point Core inflation lost ability to track underlying inflationary pressure. Thailand was the last country that use core inflation as a policy target 3

4 Administered Prices 4

5 Thailand s share of administered price has always been high above 30% of the CPI Basket Among inflation targeting countries, Thailand has the largest degree of price subsidies with 135 items, they accounted for around 35% of 2011 CPI Basket 40% 35% 30% 25% 20% 15% 10% 5% 0% Share of administered prices in CPI Basket 38% 30% 21% Average : 17% 16% 16% 17% 18% 19% 9% 10% 10% 5% Development of Administered Prices in Thailand Weight in Headline Weight in Core Number of Items Source: BIS (2009) Source: Bank of Thailand (2014) 5

6 How could the sizable share of administered prices impact monetary policy under inflation targeting? High degree of price controls Maintaining people cost of living Distort market mechanism and do not provide long term alleviation Limiting prices fluctuation and provide short term relief As control is altered, the surprising shock would drive up inflation greatly Coping with temporary shock when monetary policy cannot accommodate Market does not reflect the real pricing, and inflict costs on the economy 6

7 A consistently high proportion of government price control since 2002 could be considered as a special challenge for inflation targeting framework Administered prices also set a special challenge for the inflation targeting framework. The extra measures introduced in 2009 were the main reason for the inflation target being missed in that year. Administered prices also make it harder for the BoT to assess the current state of demand. By keeping prices down, pressures on the index may be suppressed, but at the same time holding prices down gives the public more to spend on other things. - Grenville and Ito (2010, p.44) Controls do not provide long term relief from inflation and do inflict costs on the economy by distorting the pricing system. - Daniel (1975) Large portion of administered items in CPI of emerging economies have substantial impact on inflation volatility in the short term. - Masson et al. (1998) and Debelle, (2001) Price controls caused inflation to diverge from normal trend. - Mallikamas and Ponsaparn (2005) 7

8 Research Questions, Method Strategies, and Data Construction Research questions Methodology Data How does administered price impact inflation targeting? I. Dynamics Correlation Cross Correlation (Lead Lag) Johansen Co-integration Test Auto-regression II. Transmission Mechanism VAR with impulse shock on policy rate III. Monetary Policy Response Taylor Rule CPI Index : with disaggregate models of administered prices index and marketdetermined index are constructed for analyses Core Core Administered Prices Core Free Prices Headline Headline Administered Prices Headline Free Prices *Thanks to the BOT for the CPI index dataset *Note: calculation method is demonstrated in section 3.1 of the report 8

9 Containing large proportion of administered prices items, headline inflation is more volatile Asian Financial Crisis Oil shock Euro Crisis 0-5 Global Financial Crisis Headline Headline admin Headline free Source: Bank of Thailand (2014) 9

10 I. Dynamics Coefficient Correlation Correlation of key price measures between 1994 and 2014 Core Core Admin Core Free Headline Headline Admin Headline Free Core Core Admin Core Free Headline Headline Admin Headline Free Headline inflation is strongly correlated with administered price inflation than free price inflation. This suggests a challenge on using monetary policy to control headline inflation. 10

11 Cross Correlation (Lead and Lag relationship) Headline administered price inflation and free price inflation Sample: 2000Q3 2014Q4 Included observations: 58 Correlations are asymptotically consistent approximations H_ADINF,H_FINF(-i) H_ADINF,H_FINF(+i) i lag lead Tight correlation in lag and lead periods between the twos Government authorities suppressed the prices of administered items The inflationary shock pass through to free price inflation 11

12 Johansen Co-Integration Test Administered prices do not have significant effect on comprehensiveness and do not reflect the underlying cost of living in the long term Persistency (Auto-regression) π t = μ + ρπ t 1 + ε t Estimates of persistence in disaggregate inflation ρ coefficient Prob. Core Core Admin Core Free Headline Headline Admin Headline Free *Note: Quarterly Data (%QoQ) π t = (1 γ)π t 1 + γe t π t+1 + v t Given lower persistence in headline inflation, BoT has higher ability to anchor inflation expectation 12

13 II. Monetary Transmission Mechanism Vector Auto-Regression Model Given impulse shock on policy rate, headline inflation as well as core inflation tend to adjust faster to interest rate shock than the marketdetermined price inflation does Response of headline inflation to policy rate Response of headline inflation to policy rate 8 quarters Response of free price inflation to policy rate Response of headline free prices inflation to policy rate Bank of Thailand should even be more forward looking as monetary policy likely takes longer horizon than expect Note: *Model already taken into account of exogenous oil price, price puzzle still exist quarters

14 III. Monetary Policy Response Taylor Equation Estimation Results i t = α + β 1 π t π + β 2 y t y + ρi t 1 + ε t Sample period (2003Q1-2014Q4) Normal Equation Estimated by GMM Smoothing Equation Estimated by GMM Regression regression recovered constant α = r f + π α (0.214)** (0.224)** r f + π output gap β 2 = β y β (0.116) 2 (0.049)* β y inflation gap β β (0.187)** 1 (0.101)** β π lagged policy ρ (0.079)** ρ R-squared Notes: Brackets are standard errors. ** denotes significance level at 5% Coefficients of inflation deviation in both equation are consistent with the Taylor principle (i.e. optimal monetary policy). 14

15 Monetary policy does not appear to be optimal given that it responses to jump in administered price Modified Taylor Rule with Administered Price i t = α + β 1 π admin π + β 2 π free π + β 3 y t y + ρi t 1 + ε t Smoothing Equation Estimated by GMM core inflation headline inflation Constant r f + π output gap β y administered price inflation gap β π admin free price inflation gap β π fr ee lagged policy ρ R-squared Note: *Different specifications might be needed for robustness check Central bank seems to be slightly more responsive to administered price deviation. Outcome-based Taylor rule specification is that operational policy setting is not optimal when looking at market-determined inflation alone. 15

16 Conclusion Evolving Process of Dynamics Transmission Mechanism Monetary Policy Response Price controls raise concern on controllability of Monetary Policy process has largely been influenced by price controls Strong correlation with administered prices sector Complicated lead and lag relationship No long-term relationship between price controls and inflation Headline inflation is less persistent; better anchor expectation Takes longer time for Monetary Policy to become fully effective Free prices sector has slower transmission process than aggregate headline inflation with over 2 years Optimality of Monetary Policy depends crucially on the measures of inflation Taylor Rule holds when aggregate inflation is used to estimate Reaction function seems to exhibit higher degree of response to deviation in administered price sector 16

17 Key Takeaways 1 Price controls set special challenges for monetary policy and central bank should incorporate this fact 2 3 Bank of Thailand should be even more forward looking in setting appropriate monetary policy Coordination between Bank of Thailand and Ministry of Commerce is necessary required 4 Further study: Welfare analysis on the optimal threshold for the degree of government price control (c.f. Safouane et.al. 2012) 17

18 THANK YOU. 18