نقش اوراق قرضه دولتی تحت فشار در تحول بزرگ سیاست پولی آمریکا

A fundamental shift in monetary policy occurred around 1980: the Fed went from a “passive” policy to an “active” policy. We study a model in which government bonds provide transactions services. We present two calibrations of our model, using pre- and post-1980 data. We show that estimates of pre- and post-1980 policy rules all lie within our determinacy regions. But, the pre-1980 policy was a very bad monetary policy, even if it avoided sunspot equilibria. Model simulations suggest that household welfare would have increased by 3.3 percent of permanent consumption in this period under an active policy.

A profound change in monetary policy is generally thought to have occurred in the United States around 1980: the Fed became much more responsive to inflation with the arrival of Paul Volcker. The pre-Volcker policy has been strongly criticized by Clarida et al. (2000) and many others who argued that the policy violated the Taylor principle, raising the specter of in-determinacy and sunspot equilibria, and indeed inflation was volatile during the 1970s. However, Leeper (1991) and others argued that indeterminacy problems need not arise when monetary policy violates the Taylor principle if that monetary policy is accompanied by an appropriate fiscal policy; that is, the Fiscal Theory of the Price Level (FTPL) can explain away the possibility of sunspot equilibria in the 1970s. Here, we offer an alternative to both sunspots and the FTPL. We assume that government bonds have a liquidity value and are an imperfect substitute for money, and we show that this fact in and of itself can rule out sunspot equilibria during the pre-Volcker era. There is therefore no need to resort to the FTPL if one doubts the likelihood of sunspot equilibria. Our approach has several advantages over the FTPL as an explanation of the inflationary 1970s. In addition to avoiding the many controversies surrounding the FTPL, our approach eliminates a thorny coordination problem that arises with the FTPL’s explanation.1 But does that mean that Fed’s policy in the pre-Volcker period has gotten a bum rap, that it was not so bad after all? Our model suggests that the answer to that question is emphatically no. The weak response to inflation may not have led to sunspot equilibria, but it was prone to in-flationary responses to the supply shocks of the 1970s, and according to our model, it was disastrous for welfare. Our simulations suggest that a stronger policy would have raised household welfare by the equivalent of three and a third percent of permanent consumption during the pre-Volcker era. It is helpful at this point to be more specific about the theoretical and empirical literature that provides the background for our paper. Using Leeper’s (1991) terminology, Fed’s interest rate policy was “passive” – or violated the Taylor Principle – prior to 1980; then it became “active” – or satisfied the Taylor Principle – during the Volcker–Greenspan era.2Clarida et al. (2000) and Lubik and Schorfheide (2004) documented this shift in monetary policy and expressed the conventional view of its import: the pre-Volcker policy was bad monetary policy because conventional models implied that the price level would not be pinned down, raising the specter of sunspot equilibria. The FTPL challenged the conventional view’s assumption (often implicit) that fiscal policy is “passive”; that is, when the public debt increases, the primary surplus rises by enough to stabilize debt dynamics. Leeper (1991) raised the possibility that fiscal policy might have been “active” in the pre-Volcker era; that is, the response might have been insufficient to stabilize debt dynamics. He (and the literature that followed) studied the combinations of monetary and fiscal policies that would achieve a unique, locally stable, solution for inflation. Generally speaking, an inactive monetary policy has to be paired with an active fiscal policy, and an active monetary policy has to be paired with an inactive fiscal policy.3 Pairing a passive monetary policy with a passive fiscal policy – the conventional view of the pre-Volcker period – would generally lead to sunspot equilibria. On the other hand, if the passive monetary policy was paired with an active fiscal policy, then the sunspot equilibria would be eliminated. But, when monetary policy shifted from passive to active around 1980, fiscal policy would have had to shift from active to passive, or explosive solutions would have been possible. This is the FTPL’s coordination problem.4 In Canzoneri and Diba (2005), we showed that both active and passive monetary policies can achieve price determinacy, even if fiscal policy is passive, as long as government bonds provide transactions services.5 Fiscal deficits increase total transactions balances, providing a new link between inflation and debt dynamics.6 This dramatically changes the combinations of monetary and fiscal policies that result in a unique, locally stable, equilibrium. In this paper, we complete the analysis of the 1980 transformation in monetary policy. We present two calibrations of our model, using pre- and post-1980 datasets, and we illustrate the range of monetary and fiscal policies that would have achieved price determinacy. Lubik and Schorfheide (2004) estimates of the pre- and post-1980 interest rate rules and Bohn, 1998 and Bohn, 2004 estimates of passive fiscal reaction functions fall within our determinacy regions. This is why we say that there appears to have been no need for fiscal policy to have shifted in 1980. Before we begin our analysis, we should mention some alternative views and approaches to the issues we discuss here. Orphanides (2004) finds that monetary policy in the pre-Volcker period did not actually violate the Taylor Principle; if one accepts his results, there is no indeterminacy problem to be explained. Galí et al. (2004) and Bilbiie and Straub (2006) add “rule of thumb” households to their models and find that, if there are enough of them, the Taylor Principle can be violated without creating an indeterminacy problem.7Bilbiie and Straub (2006) find that a passive monetary policy is optimal in their model when there are enough “rule of thumb” households. Woodford (2001), using the FTPL, argues that an interest rate peg is a good characterization of US policy between 1942 and the Treasury-Fed “Accord” of 1951, and he seems to think that this was a reasonable policy at that time; he says that “This sort of relationship between a central bank and the treasury is not uncommon in wartime, …[and in other] cases where the perceived constraints on fiscal policy have been similarly severe.” The rest of the paper is organized as follows: In Section 2, we outline our model, and we present our two calibrations of it: one to the 1970s and the other to the Volcker–Greenspan era. In Section 3, we graph the policy regions for stable equilibria, sunspot equilibria and explosive solutions under the two calibrations. The regions have shifted over time, but as noted above they suggest that the systematic part of fiscal policy need not have changed when monetary policy shifted from passive to active. In Section 4, we argue that the passive policy of the pre-Volcker period was indeed bad monetary policy, even if it did avoid sunspot equilibria. Finally, Section 5 concludes with some caveats and some suggestions for future research.

In this paper, we modified a standard NNS model by letting government bonds be an imperfect substitute for money. This modification brings a new – and very direct – mechanism for debt dynamics to feed into inflation dynamics: deficits increase “effective” transactions balances, and this affects inflation. The probable existence of this mechanism has a number of important implications for NNS modeling. It alters our views on the coordination of monetary and fiscal policy that is required to provide a stable nominal anchor. The Taylor Principle is no longer sacrosanct; and breaking with the FTPL, a weak fiscal response to debt is no longer the panacea for a weak monetary policy. These basic notions were already known from the work of Canzoneri and Diba (2005) and Linnemann and Schabert, forthcoming and Linnemann and Schabert, 2009. Here, we concentrated on a fundamental transformation in US monetary policy that occurred around 1980. We calibrated our model to pre- and post-1980 data and showed how, in policy parameter space, the regions for determinacy, sunspots and explosive solutions have shifted over time. But typical estimates of interest rate policy rules before and after the break, combined with typical estimates of fiscal reaction functions (all passive), all fall in the determinacy region for both calibrations. This suggests that there was no need for a shift in fiscal policy around 1980. In this respect, our interpretation of the transformation of monetary policy may be more plausible than FTPL’s: the FTPL interpretation requires an accompanying change in fiscal policy, and this coordination of monetary and fiscal policies has not been documented.18 We also argued that the pre-Volcker policy was bad monetary policy, even if it did not raise the specter of sunspot equilibria. Model simulations showed that simply making the passive policy active would increase household welfare by over three percent of permanent consumption. This welfare gain dwarfs the welfare differences between active policy rules. Finally, we showed that the passive policy of the pre-Volcker period allowed a government spending shock to pass more forcefully to output than an active policy would. This in and of itself is not very surprising: an active policy raises real interest rates, choking off the effect on aggregate demand. However, the result does suggest that our model might shed light on a transformation in the government spending multipliers that is also thought to have occurred around 1980. Perotti’s (2004) VAR analysis shows that the multipliers were strong (in the US anyway) prior to 1980, and weak to negligible after 1980.19 To link the transition in monetary policy to the transition in government multipliers in a convincing way, however, we would have to model investment. We leave this for future work.