Nonlinear and uncertainty effects of macroeconomic policies
Progetto di rilevante interesse nazionale (PRIN) 2017
Abstract
The aim of the present project is to study the nonlinear effects of monetary and fiscal policy by using a simple extension of standard VAR methods. In particular, we shall focus on four key policy issues: (1) Are the effects of monetary policy asymmetric? (2) Is the fiscal stimulus more effective during economic downturns? (3) Do monetary and fiscal shocks interact in some way? (4) Does macroeconomic policy
Question (1) has a long history in economics. The idea is that reducing interest rates might be ineffective to stimulate recovery. There are a few empirical works aimed at verifying empirically whether monetary policy effects are asymmetric. Existing evidence is generally in favour of asymmetry. However, the methods used are not fully satisfactory because of questionable identification of the policy shock or questionable technical assumptions. A confirmation obtained with a different a simple method would then be very useful.
The empirical debate about Question (2) has been originated by Auerbach and Gorodnichenko (2012), who find that the fiscal multiplier is larger during economic crises, providing a strong empirical support to Keynesian policy recommendations. Existing evidence is mixed: a few papers confirm the result, but the authoritative work by Ramey and Zubairy (2017) does not.
About Question (3) we have a large theoretical literature and some empirical work, which however focus mainly on the effects of different combinations of systematic policy rules. To our knowledge, there are no works studying nonlinear interactions of discretionary shocks. Finally, despite the large and growing literature about uncertainty measurement and the effects of uncertainty on economic activity, there is still little work about how to measure monetary and fiscal policy uncertainty and their macroeconomic effects. In particular, the announcement effects of fiscal policy may be ineffective if there is large uncertainty about their size and timing.
All of the above questions may have major policy implications. If monetary policy is ineffective to foster recovery, it should be used in combination with the fiscal stimulus, in line with the Keynesian stance. The Keynesian view would obtain further support from a positive answer to Question (2). Interactions between monetary and fiscal shocks would provide additional motivation for a coordination of fiscal and monetary policy. In the Euro area, coordination requires greater tax and spending integration. Lastly, empirical results about Question (4) may provide useful indications on the way in which fiscal policy measures should be formulated and announced.
All of the above questions call for a representation of the macro economy where structural shocks affect macroeconomic variables via nonlinear impulse-response functions. Questions (1), (3) and (4) cannot be addressed by standard nonlinear models. The method recently proposed by Matthes and Barnichon (2015) is well suited, but requires the Gaussianity assumption, which is typically rejected by macroeconomic data. In this project we shall use a new method, which is a straightforward extension of standard structural VAR methods.
The advantage of our proposed method is twofold: first, it does not require any assumption on the density distribution of the shocks; second, it is simple, in that it requires only ordinary least square estimation. The estimated policy shocks are the same as those of standard monetary and fiscal VAR models, which helps comparison of the impulse response functions. In addition, standard VARs are special cases, so that standard methods for nested models can be used to test for nonlinearity.
Durata: 2019 - 2022
Finanziamento:
Principal investigator: Mario Forni
Componenti gruppo di ricerca (UniMoRe): Luigi, Brighi, Barbara Pistoresi, Antonio,Ribba
Partners: Università degli Studi di Torino, Università Commerciale "Luigi Bocconi" Milano
[Ultimo aggiornamento: 15/05/2020 11:23:25]