PDF《RESEARCH DIVISION》

(b) in?ation is strongly correlated across countries and the correlation is stronger than that of output;

and (c) money growth is not signi…cantly correlated across countries. We then show that models with nominal rigidities and independent monetary shocks cannot simultaneously explain these stylized facts. Since the literature has already found that the sticky-information model is a good model to explain fact (a), we focus …rst on the sticky-information model'

s ability to explain fact (b). After showing that the sticky-information model cannot explain fact (b) without deteriorating its ability to explain fact (a), we then show that the same is true for sticky-price models as well.

2 For an example of solving DSGE models with sticky information by traditional methods, see Andrés, López- Salido, and Nelson (2005). The literature studying sticky-information models includes Andrés, López-Salido, and Nelson (2005), Ball, Mankiw, and Reis (2005), Coibion (2006), Jensen (2005), Keen (2005), Koenig (2004), Mankiw and Reis (2002), Mankiw, Reis, and Wolfers (2003), and Trabandt (2005), among others.

3 2 Stylized Facts We begin by looking at postwar in?ation dynamics for

18 developed countries. The data are from the IMF …nancial statistics and the OECD database (see Appendix A for detailed data descriptions). The in?ation rates are computed based on the consumer price index (CPI).3 We use the Band-Pass …lter to de…ne the output gap (business cycle components of GDP). The results are similar if the HP …lter is used instead. Some OECD countries have only seasonally unadjusted data available for prices and money stocks. In that case, seasonal adjustment by the X-11 …lter is performed whenever needed. Figure

1 graphs the domestic relationships (correlations) between output (yt) and in?ation ( t) for each of the individual countries in our sample, for leads and lags up to six quarters. Several patterns emerge from the graphs. First, the contemporaneous correlation between output and in?ation is positive for all individual countries except Norway and Portugal, where the correlation is zero. Second, in?ation systematically lags output. The maximum correlations betwe........