The hard road to a soft landing: Evidence from a (modestly) nonlinear structural model

Post‐COVID inflation dynamics: Higher for longer

“…This lengthy return to the inflation target is consistent with the inflation experience over the 2012-2019 period, when trend inflation moved mere 0.5 ppts. We further note that our core PCE projections are very similar to those of the model in Verbrugge and Zaman (2023a), which is built upon trimmed-mean PCE inflation, and also to those from the headline PCE forecasts of the model in Verbrugge and Zaman (2023b).…”

“…Regarding that persistence, allowing for (though not imposing) the fifth lag in each of the three core PCE component inflation variables is quite important. An autoregressive process with a weight of (say) 0.8 on the first lag, 0.1 on the fifth lag, and 0 on all other lags is more persistent than an autoregressive process with a weight on 0.9 on the first lag and 0 on all other lags (see appendix of Verbrugge & Zaman, 2023a). The model forecasts imply that it takes a very long time for inflation to return to trend.…”

“…In each equation, we allow up to five quarterly lags in the dependent variable and up to four quarterly lags in each of the other variables. (Allowing for the fifth lag is quite important for accurately assessing the persistence of each series, as demonstrated in Verbrugge & Zaman, 2023a. ) The equations below reflect our final preferred specifications for each equation.…”

“…20 The forecasts from our model are competitive with most alternatives post-2019; given the unusual shocks over this period, few models performed well. An exception is Verbrugge and Zaman (2023a). 21 For unemployment, there is no standard alternative benchmark, so we also report the forecast accuracy of an AR(1) model.…”

“…1 Importantly, we carefully model the relationship between unemployment and these inflation components. Recent research (Ashley & Verbrugge, 2023;Verbrugge & Zaman, 2023a) has demonstrated benefits to specifying the relationship between aggregate inflation and the unemployment rate as frequency-dependentthat is, to using a specification that allows inflation to respond differently to business-cycle movements in the unemployment rate than it does to low-frequency movements or to transient movements. 2 This previous work has established that the response of aggregate inflation to the various frequency components of the unemployment rate gap (i.e., to the persistent, moderately persistent, and transient components) varies not only by persistence component but also by the sign of these components.…”