Read an Excerpt
NBER Macroeconomics Annual Volume 28 (2014)
By Jonathan A. Parker, Michael Woodford The University of Chicago Press
Copyright © 2014 The University of Chicago
All rights reserved.
ISBN: 978-0-226-16540-0
CHAPTER 1
Dormant Shocks and Fiscal Virtue
Francesco Bianchi, Duke University and CEPR
Leonardo Melosi, Federal Reserve Bank of Chicago
I. Introduction
The importance of modeling the interaction between fiscal and monetary policies goes back to the seminal contribution of Sargent and Wallace (1981). However, in many of the models that are routinely used to investigate the sources of macroeconomic fluctuations, fiscal policy plays only a marginal role. The vast majority of papers resolve the problem of monetary/fiscal policy coordination assuming that the fiscal authority stands ready to accommodate the behavior of the monetary authority, keeping the process for debt on a stable path. This is a strong assumption as a casual observation of the data shows that countries often experience prolonged periods of severe fiscal imbalance. Quite interestingly, these episodes are frequently followed by significant increases in inflation. In some cases, such increases are short lasting and remarkably violent. In other cases, they unfold over many years, generally starting small and then gaining momentum. In this paper, we develop a theoretical framework that can quantitatively account for persistent and accelerating increases in inflation and for the heterogeneity, across countries and over time, of the link between inflation and fiscal discipline.
We model an economy populated by a continuum of agents that are fully rational and understand that debt can be stabilized through movements in taxes or movements in inflation. When the fiscal authority is virtuous and moves primary surpluses in response to fluctuations in the ratio of debt to gross domestic product (GDP), the Central Bank has full control over inflation. Under the assumption of nondistortionary taxation, fiscal shocks do not have any effect on the real economy as they only redistribute the timing of taxation. When policymakers deviate from the virtuous regime, with the fiscal authority not reacting to debt fluctuations and the Central Bank disregarding the Taylor principle, two situations can arise. If agents expect the return to the virtuous regime to be close enough in time, inflation stability is preserved. On the other hand, if the deviation is expected to last for a long period of time, high levels of debt require an increase in inflation.
We build on this basic intuition and assume that when facing a deviation from the virtuous rule, agents do not know how long it will take to move back. Instead, they have to conduct Bayesian learning to infer the nature of the deviation. As they observe more and more deviations, they get increasingly convinced that a prompt return to the virtuous regime is very unlikely. Given that agents are fully rational and understand that debt has to be financed in one way or the other, the drift in agents' beliefs determines a progressive increase in inflation. The initial movement can be almost undetectable, but as initially optimistic agents become relatively pessimistic, inflation accelerates, gaining momentum and getting out of control. At the same time, expected and realized volatilities go up as shocks that are dormant under the virtuous regime slowly start manifesting themselves. Therefore, if an external observer were monitoring the economy focusing exclusively on output and inflation, he would detect a run-up in inflation and an increase in volatility without any apparent explanation. The observer might then conclude that the volatility of the exogenous shocks and the target for inflation have both increased.
Dormant shocks are undetectable when policymakers are virtuous or agents are optimistic that they will be virtuous in the future because agents understand that any imbalance in the debt-to-GDP ratio will be followed by a fiscal adjustment. As agents become discouraged about policymakers' future behavior, the effects of dormant shocks arise. Therefore, dormant shocks can have effects many years after they occurred, as long as the fiscal imbalance that they generated is not totally reabsorbed by the time the deviation from the virtuous regime takes place. Furthermore, even after a regime change, their effects can barely be detected if agents find it extremely unlikely that policymakers will engage in a long-lasting deviation from the virtuous regime. In other words, depending on policymakers' fiscal virtue, inflation can stay low for many periods, as it takes time for agents to become convinced that the economy has entered a long-lasting deviation. According to the same logic, if on average policymakers spend a lot of time in the virtuous regime, agents might remain confident about ultimately responsible fiscal behavior even when observing a long sequence of deviations. However, no matter how optimistic agents are or how virtuous policymakers have been in the past, if a deviation lasts for an extended period of time, agents will eventually become convinced that a quick return to the virtuous regime is unlikely. In other words, following a deviation, fiscal virtue can delay the effects of dormant shocks, but it cannot eliminate them.
The interaction between dormant shocks and fiscal virtue also provides an appealing explanation for why countries with different levels of debt might have similar levels of inflation for prolonged periods of time, but then experience very different outcomes during hard times. When a virtuous regime prevails or agents are confident that it will prevail in the future, the level of debt is substantially irrelevant. However, if agents become convinced that the economy has entered a long-lasting deviation, then the differentials between the interest rate and inflation open up. The larger the difference in fiscal virtue, the larger the difference in the speed of learning, the faster the opening of the differentials between the interest rate and inflation.
Therefore, our theoretical framework is capable of accounting for the instability of the link between fiscal discipline and inflation. In our model, agents are fully rational, but uncertain about the way the trade-off between inflation and taxation will be resolved. This creates a continuum of regimes indexed according to agents' beliefs and a smooth transition from the law of motion that prevails under the virtuous regime to the one that characterizes a long-lasting deviation. Therefore, the strict distinction between Ricardian and non-Ricardian regimes typical of the fiscal theory of price level literature (Leeper 1991; Sims 1994; Woodford 1994, 1995, 2001; Schmitt-Grohe and Uribe 2000; Bassetto 2002; and Cochrane 1998, 2001, among others) breaks down and is replaced by a series of intermediate regimes that reflect the evolution of agents' expectations about the future conduct of fiscal and monetary policies.
Furthermore, agents know that they do not know. Therefore, when forming expectations, they take into account that their beliefs will evolve according to what they observe. Given this feature, our approach is clearly different from the one used in the traditional literature about learning, which assumes anticipated utility; that is, that agents form expectations conditional on their beliefs without taking into account that these are likely to change in the future. In our context, it is possible to go beyond the anticipated utility assumption because there is only a finite number of relevant beliefs and they are strictly linked to policymakers' behavior through the learning mechanism, in a way that we can keep track of their evolution.
In this respect, our paper is related to Eusepi and Preston (2012), who study the problem of macroeconomic stability in a model in which agents use adaptive learning to make forecasts about the future evolution of fiscal and monetary variables. In their model, there are not regime changes. If agents were fully rational, fiscal policy and the maturity structure of debt would be irrelevant because the Taylor principle always holds and fiscal policy is always Ricardian. However, agents do not know the parameters of the model and they erroneously believe that the economy is subject to regime changes. For this reason they use a constant gain learning algorithm in which recent observations receive more weight than observations that are far into the past. In this context, non-Ricardian effects arise because agents might erroneously regard bonds as net-wealth as in Barro (1974). Instead, in this paper non- Ricardian effects arise in the moment fully rational agents, in response to changes in policymakers' behavior, become discouraged about debt sustainability being guaranteed by movements in primary surpluses.
Given that the underlying mechanism relies on uncertainty around the source of financing for the debt-to-GDP ratio, all shocks that move this variable are potentially candidates for dormant shocks. In an environment with nondistortionary taxation, shocks to transfers and taxes are particularly interesting, given that they do not have any effect on the macroeconomic variables when the virtuous regime is in place but can generate large fluctuations in inflation once policymakers start deviating. Furthermore, given that agents are forward-looking, even announced changes in expenditure or taxation would trigger the inflationary mechanism.
We illustrate the key properties of the model using the basic three-equations new-Keynesian model used by Clarida, Gali, and Gertler (2000), Woodford (2003), and Lubik and Schorfheide (2004), augmented with a fiscal block. We then conduct a quantitative analysis building on the empirical results obtained by Bianchi and Ilut (2012). The estimates from that paper are used to calibrate a richer model and to provide guidance in characterizing the evolution of policymakers' behavior: A prolonged deviation from the virtuous regime started in the late 1950s and ended with the appointment of Paul Volcker as Federal Reserve Chairman. The transition matrix controlling the evolution of policymakers' behavior is then chosen in a way to match the stylized facts regarding the US Great Inflation: inflation started increasing in the mid-1960s, gained momentum in the early 1970s, got out of control toward the end of that same decade, and experienced a sudden drop in the early 1980s. The entire run-up of inflation of the 1970s can be obtained by considering only two shocks. The first rise of inflation would be the result of the announcement of the Great Society initiatives of President Lyndon Johnson around 1964, while the second acceleration would be caused by the tax cuts enacted by President Gerald Ford's administration. The progressive deterioration of agents' beliefs explains why inflation seemed to gain momentum over time. The appointment of Volcker at the end of the 1970s marks the return to the virtuous regime and determines the sudden drop in inflation of the early 1980s.
We then use the model to analyze the current situation. Given that dormant shocks might take a long time to unfold, we should not interpret the current low levels of inflation expectations and long-term interest rates as reflecting a low risk of high inflation for the US economy. We show that if US policymakers were to follow the current policy mix for a prolonged period of time, inflation might quickly accelerate and get out of control. In other words, the low inflation expectations and long-term interest rates reflect the reputation that US policymakers have built in the past twenty to thirty years since the Volcker disinflation. This stock of reputation is not unlimited, and it slowly deteriorates as policymakers keep deviating. This also suggests that if inflation is the result of a lack of fiscal discipline, central bankers cannot simply wait to see inflation in order to decide to worry about that. At that point, only an immediate change in both fiscal and monetary policies would be able to cut the inflation spiral.
This paper is part of a broader research agenda that aims at understanding the role of fiscal policy in explaining changes in the reduced form properties of the macroeconomy. In this regard, the current paper is related to Bianchi and Ilut (2012), who estimate a dynamic stochastic general equilibrium (DSGE) model for the US economy that allows for a structural break from a non- Ricardian regime to a Ricardian one. The main contribution of that paper is to identify the timing of the structural break, occurring a few quarters after the appointment of Volcker, and to show that the policy change can account for the rise and fall of inflation and the changes in the reduced form properties of the macroeconomy. The current paper does not present a fully specified estimation exercise. Instead, we use the results obtained by Bianchi and Ilut (2012) to fix parameter values and the timing of regime changes. That said, the current paper contributes to this research agenda in several ways. First, it introduces the notion of dormant shocks. This resolves an apparent puzzle of the fiscal theory of price level; that is, the fact that in the data fiscal shocks do not seem to cause an immediate increase in inflation. Instead, thanks to the learning mechanism, dormant shocks can cause accelerating and persistent increases in inflation that unfold over decades. Second, it illustrates how the interaction between dormant shocks and fiscal virtue can account for the heterogeneity across countries in the link between fiscal discipline and inflation. Finally, it puts the theory to work to discuss its implications for the future behavior of inflation.
Our paper is related to the extensive literature that explores the evolution of output and inflation over the past sixty years using micro-founded models. Fernandez-Villaverde, Guerron-Quintana, and Rubio-Ramirez (2010) consider models with time-varying structural parameters and find substantial evidence of parameter instability. Using a large-scale DSGE model augmented with stochastic volatilities, Justiniano and Primiceri (2008) find that changes in the volatility of investment shocks play a key role in explaining the evolution of the reduced form properties of the economy. Davig and Leeper (2007), Bianchi (2013), and Davig and Doh (2008) allow for heteroskedasticity and changes in monetary policy. Bianchi and Melosi (2012a) develop a theoretical framework to quantitatively assess the general equilibrium effects and welfare implications of central bank reputation and transparency. Finally, Ireland (2007), Liu, Waggoner, and Zha (2011), and Schorfheide (2005) consider models in which the target for inflation is moving over time. Our model is able to account for changes in the low-frequency component of inflation and in the volatility of the endogenous variables.
Our work is also related to Benati (2008); Cogley, Primiceri, and Sargent (2010); Cogley, Sargent, and Surico (2011); and Coibion and Gorodichenko (2011). Benati (2008) documents that inflation persistence is not stable across time and across countries. Cogley et al. (2010) study changes in the persistence of the inflation gap measured in terms of short- to medium-term predictability. Cogley et al. (2011) show that the Gibson's paradox (i.e., low correlation between inflation and nominal interest rates) vanished during the Great Inflation and reappeared after 1995. Coibion and Gorodichenko (2011) point out that the determinacy region in a model with positive trend inflation could be smaller than what is implied by the Taylor principle. They conclude that the US economy was still at risk of indeterminacy in the 1970s, even if the Taylor principle was likely to be satisfied, because of the high level of trend inflation. Our model is able to generate variability in the persistence and low frequency component of inflation as a result of the evolution of agents' beliefs about policymakers' future behavior. Finally, our work is also linked to papers that study the impact of monetary policy decisions on inflation and inflation expectations, such as Mankiw, Reis, and Wolfers (2004); Nimark (2008); Del Negro and Eusepi (2011); and Melosi (2012, 2013).
This paper can be summarized as follows. In section II, we describe the model, outlining its properties under fixed coefficients. In section III, we introduce regime changes and learning. In section IV, we introduce the notion of dormant shocks and explain how they are related to fiscal virtue. We put the theory to work in section V: first, we look at the past; and then we look at the current situation and beyond under the same theoretical framework. We present our conclusions in section VI.
II. The Model
In order to illustrate the key properties of the model, we consider the basic new-Keynesian model employed by Clarida, Gali, and Gertler (2000), Woodford (2003), and Lubik and Schorfheide (2004), augmented with a fiscal rule. This model has very little built-in persistence, given that it features a purely forward-looking Phillips curve. This will allow us to isolate the effects of the learning mechanism.
(Continues...)
Excerpted from NBER Macroeconomics Annual Volume 28 (2014) by Jonathan A. Parker, Michael Woodford. Copyright © 2014 The University of Chicago. Excerpted by permission of The University of Chicago Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
