We choose a flexible trimmed mean a

We choose a flexible trimmed mean as our measure of underlying inflation. Trimmed-mean inflation removes extreme movements in prices compared to the general tendency, thus mitigating effects of a cross-sectional price growth distribution with thick tails. The percentiles of the cross-section distribution of price changes to be eliminated are determined by the coherence of the resulting underlying inflation measure with the output cycle. With the exception of Vega and Wynne (2001), analysis using the trimmed mean inflation measure for the euro area is scarce.
There has been more research on the output-inflation comovement for the US economy than for the euro area. The Phillips correlation has been doubted, see e.g. Cooley and Ohanian (1991), but also argued for numerous times, see e.g. King and Watson (1994), Sargent (2001) or Stock and Watson (2010) whose sample cover the Great Recession. Recently Andrle (2012) demonstrates a strong, stable and positive comovement between the cyclical component of output and deviation of core inflation from long-term inflation expectations in the U.S. starting from the 1960’s through the Great Recession. This paper documents strong comovement of output and inflation using a set of nonrestrictive assumptions, rather than constructing a formal semi-structural model of the Phillips curve, with many restrictive assumptions as in Basistha and Nelson (2007) for instance.5 Recently, Basturk and others (2013) proposed