了解商业周期各阶段的通货膨胀与产出关系

Understanding the inflation–output Dario Cardamone, Roberto A. De Santis of the Phillips and DIS curves, the Federal Reserve’s policy response, and the impact of economic shocks differ across economic states.Our findings reveal significant differences across regimes.The Taylor rule, which target and output is above potential, while reactions to inflation is broadly similar acrossregimes. We also show that the Federal Reserve is more data dependent during periodsof inflationary slack. Moreover, we find that the size of monetary policy shocks is signif-icantly larger when inflation exceeds its target, reflecting the Federal Reserve’s attention during inflationary booms, restrictive monetary policy can effectively reduce inflationand close the output gap, while during inflationary slack, policy changes may have amore limited impact. This might imply a stronger determination to fight inflation. resource slack or inflationary pressure–is therefore central to understanding how shocks are transmitted through the economy and how monetary policy should be optimallydesigned across regimes.Adopting the Federal Reserve’s perspective, this paper investigates, within a uni-fied empirical framework, whether the slopes of the Phillips curve and the DynamicInvestment-Savings (DIS) curve, the parameters governing the monetary policy rule,and the transmission of macroeconomic shocks exhibit state dependence. We define the Our analysis further reveals that the sensitivity of the output gap to the interest rate inthe DIS curve is flatter whenπt>π∗tandyt≤0. Consequently, monetary policy shockshave muted responses on the macroeconomy only in the inflationary slack regime, which account for44% of the sample. This might imply a stronger determination by the FederalReserve to fight inflation in this regime.These muted responses are similar to thoseuncovered by Baumeister and Hamilton (2018) in their linear model.Prior empirical research studied primarily the time variation of parameter estimates betti,2019). Others found that the sensitivity of inflation to shocks diminished over the observation is consistent with the argument put forth by Harding et al. (2023) and Karadiet al. (2024).There are a limited number of studies that estimate the Euler equation for output,and these typically rely on structural single-equation models. Attanasio and Low (2004)estimate the elasticity of intertemporal substitution to be0.7.In contrast, Fuhrer andRudebusch (2004) and Cecchetti et al. (2023) find a very low elasticity of output or un-employment in response to changes in the interest rate.The range of our estimates,which varies according to the state of the economy, encompasses the elasticity estimated tion TVAR model. Section3details the empirical estimation strategy. Section4presentsthe empirical results. Section5concludes. 2A three-equation threshold VAR πt=¯κhτ−1+φyt+ξsti+βEtπt+1 withπtdenoting price inflation,ytthe output gap, andrtthe nominal policy rate. Thefirst equation expresses current inflation as a function of the output gap, expected futureinflation, and supply-side shocksξs, with the strength of these relationships determined flation responds positively to its expected future value and the output gap. Similarly, the The model is closed with a Taylor-type interest rate rule: indicators, as the policy rate is less constrained by its past values allowing parameters and shock variances depend on a given state of the economyS: rtSy,Sytπ,Sπtwherext−1=y′t−1,y′t−2, . . . ,y′t−m, 1
′, withust,uand monetary policy shocks, respectively. The state-dependent matrix of contemporane-ous coefficients,AS, is defined byAS=1−απ,S1−βπ,S ECB Working Paper Series No 3175 This formulation can be equivalently expressed in terms of structural coefficients.2 such nonlinearities, we modify the monetary policy rule as follows: Federal Reserve’s long-run inflation target is invariant across economic states.3we posit that the Federal Reserve’s responsiveness to inflation and output fluctuationsvaries with economic conditions.This framework suggests that, in certain states, theFederal Reserve may tolerate higher inflation to mitigate declines in output, consistent 3Model estimation The TVAR in (4) is estimated using Bayesian methods and is expressed as: 3.1 Priors on the contemporaneous structural coefficients Prior information on the distributions of the contemporaneous coefficients are taken with modes of0.5and1.5, respectively.4Both distributions have a scale parameter of0.4, have3degrees of freedom, and are truncated to be positive.Last, the smoothingparameterρis assumed to follow a Beta distribution with a mean of0.5and a standarddeviation of0.2, as in Lubik and Schorfheide (2004), Del Negro and Schorfheide (2004), in the interest rate is associated with a decrease in the output gap through the aggregatedemand schedule. Additionally, we apply very loose priors with a scale parameter of0.4and3degrees of freedom (e.g. Ba