For instance, backward inflation indexation in DSGE models which is a mechanical way to make the New Keynesian model match the persistence of inflation in the US data is no longer required.

This highlights the value of partial-information estimation over full-information[15/16]

I demonstrate this bias analytically and show its policy implications using the Smets & Wouters (2007) model:
🔹 Estimating the model to match identified business cycle shocks via IRF matching resolves key issues.
🔹 Many ad hoc shocks in DSGE models become unnecessary..[14/16]

I find significant evidence of such non-business cycle long-run shocks.

Separating such non-business-cycle fluctuations is crucial for counterfactual policy analysis, as their inclusion biases parameter estimates of DSGE models. [13/16]

Structural interpretation:
1️⃣ The first shock (no long-run effects) drives positive comovement in inflation & output—similar to demand shocks.
2️⃣ The second shock (both effects) drives negative comovement in inflation & output—consistent with supply or productivity shocks.[11/16]

The two identified shocks explain 99% of business cycle fluctuations of GDP, generating comovements without relying on ex-ante structural restrictions..[10/16]

Test summarized by the scree plot using the benchmark VAR of ACD, show that business cycle fluctuations are best described as having a two-factor structure: procedures that impose a single source of business cycles risk conflate the effects of two distinct structural shocks[8/16]