INFLATION EXPECTATIONS IN POLAND Tomasz Łyziak Bureau of Economic Research Economic Institute National Bank of Poland Tomasz.Lyziak@nbp.pl Workshop on: Models of Expectation Formation and the Role of the News Media for Information Transmission", Universität Hamburg June 14 t h -15 t h, 2012
Motivation Aims of the paper: to construct direct measures of inflation expectations of consumers, enterprises and financial sector analysts in Poland that would be comparable with each other; to use those measures in testing the formation of inflation expectations and to compare their characteristics across the analysed groups of economic agents. All the measures of inflation expectations are survey-based. Sample period: 2001-2011. 1/29
Content 1. Direct measures of inflation expectations applied. 2. Testing main features of inflation expectations in Poland: the degree of forward-lookingness of inflation expectations (in particular the fulfillment of the principal requirements of the rational expectations hypothesis); the impact of the National Bank of Poland inflation target on inflation expectations; the relationship between inflation expectations of different groups of economic agents and the role of the news media for information transmission; the role of inflation expectations in affecting actual price dynamics in the Polish economy. 3. Conclusions. 2/29
1. Direct measures of inflation expectations applied 3/29
Direct measures of inflation expectations Inflation expectations of all economic agents under consideration are surveybased and with a 12-month horizon. Data frequency: monthly in the case of consumers and financial sector analysts inflation expectations; quarterly in the case of producers inflation expectations. To assess the robustness of empirical findings concerning the formation of inflation expectations with respect to the measurement (quantification) techniques applied, we use more than one measure of inflation expectations of consumers and producers. In the case of quantified measures, we cross-check the results with tests using balance statistics. 4/29
Direct measures of inflation expectations Data source Consumers GfK Polonia (EC), additionally: GUS, Ipsos Data type qualitative (1) Applied measures Source of measurement uncertainty balance statistic 3 quantified probability measures measure of perceived inflation Producers NBP (Fast Monitoring) qualitative (2) (since 2008Q3) quantitative (till 2008Q2) balance statistic 2 quantified probability measures creating a single time series of inflation expectations based on quantitative data (till 2008Q2) and quantified data (since 2008Q3) Financial sector analysts Reuters quantitative quantitative volatility of the number of participants (1) Survey question: Given what is currently happening, do you believe that over the next 12 months prices will: (1) rise faster than at present, (2) rise at the same rate, (3) rise more slowly, (4) stay at their present level, (5) go down, (6) difficult to say.. (2) Survey question: In [here: the month with the most recent CPI index available] CPI inflation was % in annual terms. In your opinion during next 12 months prices will: (1) rise faster than at present, (2) rise at the same rate, (3) rise more slowly, (4) stay at their present level, (5) go down, (6) difficult to say. 5/29
Inflation expectations in Poland 9 8 7 6 5 4 3 2 1.8.7.6.5.4.3.2.1 8 7 6 5 4 3 2 1 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 current CPI inflation y/y consumer inflation expectations, objectified measure consumer inflation expectations, subjectified measure consumer inflation expectations, alternative subjectified measure producer inflation expectations, main measure producer inflation expectations, alternative measure financial sector analysts inflation expectations.0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 current CPI inflation y/y - right axis, % balance statistic of consumer inflation expectations - left axis balance statistic of producer inflation expectations - left axis 6/29
Inflation expectations in Poland Average inflation expectations in 2001-2011 were above current inflation in the case of producers and financial sector analysts and below current inflation in the case of consumers. 3.5 3 2.5 2 1.5 average consumers - objectified measure consumers - subjectified measure consumers - subjectified alternative measure producers - main measure producers - alternative measure 1 0.5 0 consumers producers financial sector analysts current inflation 7/29
2. Testing main features of inflation expectations in Poland 8/29
Expectational errors (ME) Consumers seem to underestimate future inflation, while producers and financial sector analysts overestimate it. 0.8 0.6 0.4 0.2 0 ME consumers - objectified measure consumers - subjectified measure consumers - subjectified alternative measure producers - main measure producers - alternative measure -0.2-0.4-0.6 consumers producers financial sector analysts 9/29
Expectational errors cross-check with balance statistics Balance statistic of expected inflation applied in the paper attaches the weights: 1, 0.5, 0, -0.5, -1 to the subsequent fractions of respondents (declaring that prices are expected to: increase more rapidly, increase at the same rate, increase at slower rate, stay about the same, fall ). Inflation expressed in terms of the above balance statistic can be defined in the following way: π = 1 if π π BS t t t 12 πbs t = 05. if πt= πt πbs t = 0 if 0 πt πt πbs t = 05. if πt= 0 πbs = 1 if π 0 t t 12 12 10/29
Expectational errors cross-check with balance statistics A comparison of average values of balance statistics of inflation expectations and actual future inflation confirms the results based on quantified measures of expectations consumers seem to form their opinions in excessively optimistic manner, while enterprises opinions are slightly over-pessimistic. Table. Average balance statistics of inflation expectations and actual future inflation expressed in units of the balance statistic Agents Data source, measure Average inflation expectations Average actual future inflation Difference (in % of average actual future inflation) Consumers GfK Polonia, balance statistic 0.31 0.47-34.4 Enterprises NBP, balance statistic 0.50 0.49 2.5 11/29
Expectational errors (MAE) For all the groups of economic agents under consideration inflation expectations errors lower than errors of the naive forecasts. 2.5 2 1.5 1 MAE consumers - objectified measure consumers - subjectified measure consumers - subjectified alternative measure producers - main measure producers - alternative measure 0.5 0 consumers producers financial sector analysts naive forecast 12/29
Testing rational expectations hypothesis Unbiasedness: Expectations are unbiased if they differ from actual values by random (unforcastable) residual: π = α+ β π + ε α = 0 β = 1 12 e, tt t t REH Macroeconomic efficiency: Expectations are efficient if information variables available while forming expectations are not statistically significant in explaining expectational errors: e = α + α Ω + ε REH α = t 0 1 t t 1 0 13/29
Unbiasedness Inflation expectations in Poland do not fulfill unbiasedness condition of REH. Table. Unbiasedness of inflation expectations Agents Data source, measure R 2 adj. α β Consumers Enterprises GfK Polonia, objectified 0.04 GfK Polonia, subjectified 0.05 GfK Polonia, Ipsos, GUS, subjectified alternative 0.05 NBP, main 0.02 NBP, alternative 0.02 Financial sector analysts Reuters 0.07 0.029 (0.006) 0.026 (0.033) 0.033 (0.006) 0.033 (0.003) 0.030 (0.003) 0.033 (0.005) -0.239 (0.172) -0.261 (0.160) -0.333 (0.203) -0.131 (0.107) -0.143 (0.098) -0.197 (0.136) F-prob H0: (α,β)=(0,1) 0.000 0.000 0.000 0.000 0.000 0.000 14/29
Macroeconomic efficiency The degree of macroeconomic efficiency of inflation expectations diversified across analysed groups of economic agents. Table. Macroeconomic efficiency of expectations Information variables Economic agents WIBOR1M WIBOR3M PLN/EUR PLN/USD Industrial output Unemployment rate Oil price CPI inflation Consumers - - + ++ + - - ++ - Producers ++ ++ + ++ ++ + ++ - Financial sector analysts + ++ + ++ + + ++ - - The following symbols are used: ++ given variable used efficiently while forming inflation expectations, + given variable used rather efficiently while forming inflation expectations, - - given variable used inefficiently while forming inflation expectations, - given variable used rather efficiently while forming inflation expectations. 15/29
Testing forward-lookingness of expectations We estimate two hybrid models of expectations formation: AD: a combination of rational and adaptive expectations (Gerberding 2001; Carlson and Valev 2002; Heineman and Ulrich 2006): ( 1 ) ( ) ( ) e e e π = α + α π + α π + α π π + α π π + ε t+ 12 t 1 2 t+ 12 2 t 2 t 14 3 t 2 t 14 t 2 4 t 2 t 14 t ST: a combination of rational and static expectations: ( ) e π = α + α π + + α π + 1 2 12 1 2 2+ 12 t t v t t t The degree of expectations backward-lookingness (adaptive or static expectations) proxied by the weight (1- α2), their forward-lookingness by the weight α2. 16/29
Testing forward-lookingness of expectations Polish enterprises display the highest degree of forward-lookingness. Consumer inflation expectations are mostly adaptive and the role of the forward-looking component is small. Inflation expectations of financial sector analysts display intermediate degree of forward-lookingness. 0.4 0.35 0.3 0.25 0.2 0.15 degree of forward-lookingness consumers - objectified measure consumers - subjectified measure consumers - subjectified alternative measure producers - main measure producers - alternative measure 0.1 0.05 0 consumers producers financial sector analysts 17/29
Deviations of expectations from the NBP inflation target The highest absolute deviations from the target were observed in the case of consumers, while the lowest in the case of financial sector analysts. Absolute deviations of enterprise inflation expectations from the NBP inflation target were slightly larger than deviations of financial sector analysts inflation expectations. 1.4 1.2 1 0.8 0.6 deviation from the NBP target consumers - objectified measure consumers - subjectified measure consumers - subjectified alternative measure producers - main measure producers - alternative measure 0.4 0.2 0 consumers producers financial sector analysts 18/29
How the NBP inflation target influences expectations To assess the weight to the inflation target in the formation of inflation expectations, we use the Bomfim and Rudebush (2000) approach, estimating the credibility parameter λ in the following equation: ( ) πe = λ πtar + + λ π 0 + ε + 12 t 12 1 t t t t In 1998-2003 there were annual NBP inflation targets set for the end of subsequent years, while since February 2003 there has been a continuous target 2.5%±1p.p. In order obtain implicit monthly inflation targets covering the whole period considered in this study (i.e. 2001-2011), 3 measures of inflation target are used, i.e.: the official target for a given year, a smoothened inflation target computed with the Hodrick-Prescott filter, and a monthly linear interpolation between end-year targets. 19/29
How the NBP inflation target influences expectations The degree of the NBP credibility is the highest among financial sector analysts. Polish enterprises inflation expectations display a slightly lower degree of anchoring. Polish consumers do not treat the NBP inflation target as an important benchmark in setting their inflation expectations. 0.8 0.7 0.6 0.5 0.4 0.3 weight of the NBP inflation target consumers - objectified measure consumers - subjectified measure consumers - subjectified alternative measure producers - main measure producers - alternative measure 0.2 0.1 0 consumers producers financial sector analysts 20/29
Relationship between expectations of different agents Carroll (2003) model explaining disagreement between expectations of professional and unprofessional agents (gape) with inflation media coverage (inf): gape = α + α inf + ε t 0 1 t t Estimation results show that consumer expectations in Poland are closer to financial sector agents expectations when inflation is more frequently analysed in mass media, i.e. in the periods of relatively high inflation. Near-rationality? Table. News coverage and disagreement in inflation expectations Measure of consumer inflation expectations R 2 adj. α0 α1 objectified 0.07 subjectified 0.07 subjectified alternative 0.07 1.44 (0.30) 2.04 (0.39) 1.46 (0.25) -0.84 (0.44) -1.58 (0.56) -0.58 (0.40) 21/29
Intensity of information and disagreement of expectations.07 160.06 140.05 120.04 100.03 80.02 60.01 40.00 20 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 current CPI inflation y/y absolute difference between inf. expectations of consumers (obj. measure) and financial sector analysts absolute difference between inf. expectations of consumers (subj. measure) and financial sector analysts absolute difference between inf. expectations of consumers (subj. altern. measure) and financial sector analysts absolute difference between inf. expectations of producers (main measure) and financial sector analysts intensity of information concerning inflation - right axis 22/29
Relationship between expectations of different agents Granger causality tests indicate causal linkages between inflation expectations of analysed groups of economic agents, i.e.: Financial sector analysts and enterprises inflation expectations are Granger causes of consumers inflation expectations, while consumers expectations do not influence expectations of the remaining groups. There exists two-way causality between inflation expectations of financial sector analysts and enterprises (main measure). Epidemiology models (Carroll 2003): ei ei e j t+ 4 / t 0 1 t+ 3 / t 1 2 t+ 4 / t π = β + β π + β π + ε Estimation results (quarterly data) confirm links between inflation expectations identified with Granger causality tests, indicating that there is a particularly strong long-term impact of financial sector analysts inflation expectations (main measure) on consumer inflation expectations. t 23/29
Relationship between expectations of different agents Table. Epidemiology of inflation expectations Measure of inflation expectations i j β1 β2 R 2 adj F-prob: H0: β1+β2=1 INFE_C_OB INFE_FSA 0.68 0.24 (0.08) (0.09) 0.78 0.14 INFE_C_SUB INFE_FSA 0.72 0.16 (0.09) (0.07) 0.81 0.07 INFE_C_SUB_A INFE_FSA 0.72 0.20 (0.09) (0.09) 0.81 0.28 INFE_C_OB INFE_P_M 0.63 0.27 (0.08) (0.07) 0.82 0.03 INFE_C_SUB INFE_P_M 0.68 0.18 (0.08) (0.05) 0.85 0.02 INFE_C_SUB_A INFE_P_M 0.68 0.23 (0.08) (0.07) 0.85 0.12 INFE_C_OB INFE_P_A 0.58 0.35 (0.07) (0.07) 0.84 0.08 INFE_C_SUB INFE_P_A 0.64 0.23 (0.08) (0.05) 0.87 0.02 INFE_C_SUB INFE_P_A 0.64 0.30 (0.08) (0.07) 0.87 0.24 INFE_P_M INFE_FSA 0.54 0.31 (0.17) (0.19) 0.67 0.14 INFE_FSA INFE_P_M 0.62 0.33 (0.06) (0.05) 0.92 0.00 INFE_C_OB objectified measure of consumer inflation expectations, INFE_C_SUB (INFE_C_SUB_A) subjectified (subjectified alternative) measure of consumer inflation expectations, INFE_P_M (INFE_P_A) main (alternative) measure of enterprise inflation expectations, INFE_FSA financial sector analysts inflation expectations. 24/29
Direct measures of expectations in the Phillips curve New Keynesian hybrid Phillips curve (Fuhrer and Moore 1995): π = α + α π x e t t + α π + α + α + ε t c e c r t 0 1 + 4/ 2 t 4 3 t i 4 t j Symbols: π c core inflation (two measures of core inflation are used: the first one excludes foodstuffs and fuels, the second one foodstuffs and energy), x output gap, e r real effective exchange rate Inflation expectations of all groups of economic agents under consideration are statistically significant and their impact on core inflation is stronger than the impact of past inflation. 25/29
Direct measures of expectations in the Phillips curve Table. Inflation expectations in the hybrid NKPC, core inflation measure excluding foodstuffs and fuels Agents Data source, measure, lags α1 α2 α3 α4 R 2 adj. Consumers Enterprises Financial sector analysts GfK Polonia, objectified, i=3, j=1 GfK Polonia, subjectified, i=3, j=1 GfK Polonia, subjectified altern., i=3, j=1 NBP, main, i=3, j=1 NBP, alternative, i=3, j=1 Reuters, i=3, j=1 0.62 (0.23) 0.57 (0.17) 0.45 (0.13) 1.02 (0.32) 0.78 (0.25) 1.05 (0.14) 0.20 (0.06) 0.24 (0.03) 0.24 (0.03) 0.14 (0.07) 0.22 (0.04) x 0.37 (0.19) 0.40 (0.16) 0.40 (0.16) 0.45 (0.23) 0.52 (0.19) 0.55 (0.14) -0.03 (0.01) -0.02 (0.01) -0.02 (0.01) -0.01 (0.00) -0.04 (0.02) -0.06 (0.01) F-prob H0: α1+α2=1 0.89 0.35 0.90 0.24 0.90 0.01 0.83 0.56 0.86 0.99 0.87 0.74 26/29
Direct measures of expectations in the Phillips curve Table. Inflation expectations in the hybrid NKPC, core inflation measure excluding foodstuffs and energy Agents Data source, measure, lags α1 α2 α3 α4 R 2 adj Consumers Enterprises Financial sector analysts GfK Polonia, objectified, i=3, j=1 GfK Polonia, subjectified, i=3, j=1 GfK Polonia, subjectified altern., i=3, j=1 NBP, main, i=3, j=1 NBP, alternative, i=3, j=1 Reuters, i=3, j=1 0.76 (0.15) 0.62 (0.10) 0.48 (0.08) 0.66 (0.18) 0.61 (0.16) 1.14 (0.10) 0.18 (0.06) 0.23 (0.03) 0.23 (0.03) 0.25 (0.04) 0.27 (0.02) x 0.16 (0.09) 0.26 (0.12) 0.26 (0.12) 0.48 (0.15) 0.48 (0.14) 0.42 (0.10) -0.04 (0.01) -0.03 (0.01) -0.03 (0.01) -0.04 (0.01) -0.04 (0.01) -0.06 (0.01) F-prob H0: α1+α2=1 0.91 0.59 0.89 0.09 0.89 0.00 0.86 0.59 0.89 0.45 0.87 0.15 27/29
3. Conclusions 28/29
Conclusions Survey data extremely useful in analysing inflation expectations. Results based on such measures should be checked for robustness by applying different quantification schemes. Formation of inflation expectations in Poland is diversified among analysed groups of agents, especially in terms of the efficiency with which available information is processed. There is a clear distinction between consumer inflation expectations on the one hand and inflation expectations of financial sector analysts and enterprises on the other hand. It is surprising how small (or even negligible) the differences existing between financial sector analysts and enterprises inflation expectations are. Such similarity probably results not only from capacities of enterprises in observing and forecasting changes in the macroeconomic environment on their own, but at least to some extent reflects the fact that enterprises monitor closely inflation expectations of professional forecasters and use them in setting their expectations. Differences between expectations of those groups of economic agents are inversely related to the intensity of media news concerning inflation. 29/29