Evaluación y combinación óptima de densidades predictivas procedentes de variables macroeconómicas desagregadas con especial énfasis en el análisis de la inflación

Abstract

Nowadays, many central banks operate under a regime of monetary policy called inflation targets whose main objective is price stability. In the case of the European Central Bank (ECB) is around 2 % in terms of annual rate. Although, due to the existence of so-called lags in monetary policy, in fact, it is a strategy often called inflation forecast based policy rules. Under this regime of monetary policy, the recent formulation of the New Keynesian model shows that in the design of optimal monetary policy, it plays a crucial role as the construction of optimal predictions (for example Svensson, (2005)) as the correct understanding of the economic agents about the decisions monetary policies by the central bank. Greater transparency should be a greater understanding of the public about the decisions of the Central Bank making more effective monetary policy decisions. The result is that the Central Bank published its predictions and they offer to the general public, for the example the quarterly inflation reports (Quarterly Inflation Report) elaborated by different institutions. In many of them the so-called fan chart or graph shown range. The fan chart represents the distribution function around the forecast path. That is, it represents the uncertainty associated with the point forecasts. In this sense it is essential for the decision-maker to estimate of the most possible correct way this uncertainty in order to improved the decisionmaking process. Although the estimate of the uncertainty associated with the puntual forecast has figured prominently in the empirical research, the principle of equivalence allowed to despise other features measures (measures of dispersion, shape and skewness) of uncertainty that it was not the mathematicalexpectation o the expected value. If the decision-maker has a quadratic loss function and the dynamic behavior of the economy can be adequately represented by linear equations then the principle of certainty equivalence implies that we only look at the first moment (the conditional expectation) of future values of the random variable, Simon (1956) and Theil (1957). The use of a quadratic loss function by the decision-maker implies that the preferences of the central bank are symmetrical. Works such as Karagedikli and Lees (2007), Surico (1956), Cukierman and Muscatelli (2008) found asymmetries in thepreferences of the central banks. Dolado et al. (2004) found asymmetries in the reactionfunction of monetary policy by the Federal Reserved in the Volcker-Greenspan period. Other works such as Kahneman and Tversky (1979) found similar results but from the field of psychology of choice.