H. Kaiser, The Application of Electronic Computers to Factor Analysis, Educational and Psychological Measurement, vol.20, issue.1, pp.141-151, 1960.
DOI : 10.1177/001316446002000116

G. Kapetanios and M. Marcellino, Factor-GMM estimation with large sets of possibly weak instruments, Computational Statistics & Data Analysis, vol.54, issue.11, pp.2655-2675, 2010.
DOI : 10.1016/j.csda.2010.04.008

O. Kim, S. C. Lim, and K. W. Shaw, The Inefficiency of the Mean Analyst Forecast as a Summary Forecast of Earnings, Journal of Accounting Research, vol.39, issue.2, pp.329-336, 2001.
DOI : 10.1111/1475-679X.00015

D. Kohn and B. Sack, Central Bank Talk: Does it Matter and Why?, Macroeconomics, Monetary Policy, and Financial Stability, pp.175-206, 2004.

N. G. Mankiw and R. Reis, Sticky Information versus Sticky Prices: A Proposal to Replace the New Keynesian Phillips Curve, The Quarterly Journal of Economics, vol.117, issue.4, pp.1295-1328, 2002.
DOI : 10.1162/003355302320935034

B. Mackowiak and M. Wiederholt, Optimal Sticky Prices under Rational Inattention, American Economic Review, vol.99, issue.3, pp.769-803, 2009.
DOI : 10.1257/aer.99.3.769

F. Milani, Expectations, learning and macroeconomic persistence, Journal of Monetary Economics, vol.54, issue.7, pp.2065-2082, 2007.
DOI : 10.1016/j.jmoneco.2006.11.007

S. Morris and H. S. Shin, Social Value of Public Information, American Economic Review, vol.92, issue.5, pp.1521-1534, 2002.
DOI : 10.1257/000282802762024610

G. Moscarini, Limited information capacity as a source of inertia, Journal of Economic Dynamics and Control, vol.28, issue.10, pp.2003-2035, 2004.
DOI : 10.1016/j.jedc.2003.08.002

M. Musard-gies, DO EUROPEAN CENTRAL BANK'S STATEMENTS STEER INTEREST RATES IN THE EURO ZONE?*, The Manchester School, vol.80, issue.4, pp.116-139, 2006.
DOI : 10.1016/S0304-3932(01)00055-1

R. Nunes, Inflation Dynamics: The Role of Expectations, Journal of Money, Credit and Banking, vol.40, issue.6, pp.1161-1172, 2010.
DOI : 10.1111/j.1538-4616.2010.00324.x

A. Orphanides, J. C. Williams-rosa, C. , and G. Verga, Learning, expectations formation, and the pitfalls of optimal control monetary policy, Journal of Monetary Economics, vol.55, issue.231, pp.80-96, 2007.
DOI : 10.1016/j.jmoneco.2008.08.002

C. A. Sims, Implications of rational inattention, Journal of Monetary Economics, vol.50, issue.3, pp.665-690, 2003.
DOI : 10.1016/S0304-3932(03)00029-1

J. Sturm and J. De-haan, Does central bank communication really lead to better forecasts of policy decisions? New evidence based on a Taylor rule model for the ECB, Review of World Economics, vol.19, issue.1, pp.41-58, 2011.
DOI : 10.1007/s10290-010-0076-4

J. H. Stock, J. H. Wright, and M. Yogo, A Survey of Weak Instruments and Weak Identification in Generalized Method of Moments, Journal of Business & Economic Statistics, vol.20, issue.4, pp.518-529, 2002.
DOI : 10.1198/073500102288618658

M. Woodford, Central-bank communication and policy effectiveness " , In The Greenspan era: Lessons for the future, pp.399-474, 2005.

J. Wu and F. Xia, Measuring the Macroeconomic Impact of Monetary Policy at the Zero Lower Bound, NBER Working Paper, No. 20117, 2014.
DOI : 10.1111/jmcb.12300

I. Regression, *. Residuals-on-instruments, *. *. , *. *. 10%, E. Ecb_ny et al., Standard errors in brackets while all regressors are from date t-1 . Instruments are the t-1 first 3 components of a Principal Component Analysis of ECB shadow rate, core HICP, Output gap, Credit growth, Oil prices Our main variables of interest -the three policy variables-are considered endogenous and instrumented. The equation is therefore exactly identified In column, 2, robust standard errors are estimated using heteroskedastic and autocorrelation-consistent (HAC) robust variance estimates In column 3, the ECB rate is replaced by the ECB shadow rate estimated by Wu and Xia (2014) In column 6, the GMM generalization of the LIML estimator to the case of possibly heteroskedastic and autocorrelated disturbances -the continuously updated GMM estimator or CUE-is used In column 7, the ECB qualitative communication variable MP_ST is replaced by MP_INT. In columns 8 and 9, it is the ECB quantitative communication -the ECB projections-which is replaced by the quarter-over-quarter difference in ECB projections and the difference between ECB projections and CF forecasts respectively In column 12, the ECB projections are split into 2 variables for ECB projections published during the first semester and the second one. In column 13, the ECB next year projections replace the ECB current year ones. In column 14, the frequency is quarterly The dispersion of the ECB qualitative communication is added, column 10, the sample starts in 2006m4. In column 11, the ECB projections and the ECB rate are contemporaneous to the dependent variable

. Predictor-coef, *** with Lo w M o de rato r

E. , E. Mp_st, and M. , means coefficients are significant at 10%, 5% and 1% respectively Standard errors in brackets. The dependent variable is private inflation forecasts at date t , while all regressors are from date t-1 . The interaction variable is generated from the multiplication of the predictor and the moderator variables. Instruments are the t-1 first 3 components of a Principal Component Analysis of ECB shadow rate, core HICP, Output gap, Credit growth, Oil prices, and a fourth instrument generated from the interaction of the predictor variable and the most correlated component with the moderator variable. Our main variables of interest -the three policy variables and the interaction term-are considered endogenous and instrumented. The equation is therefore exactly identified. For sake of simplicity, Table 6 -Interacting ECB communications and action All regressors are we compute the predictor coefficient while fixing the value of the moderator variable at either a high value (mean + 1 S.D.) or a low value