Methods of monetary policy implementation continue to change. The level of reserve supply---scarce, abundant, or somewhere in between---has implications for the efficiency and effectiveness of an implementation regime. The money market events of September 2019 highlight the need for an analytical framework to better understand implementation regimes. We discuss major issues relevant to the choice of an implementation regime, using a parsimonious framework and drawing from the experience in the United States since the 2007-09 financial crisis. We find that the optimal level of reserve supply likely lies somewhere between scarce and abundant reserves, thus highlighting the benefits of implementation with what could be called "ample" reserves. The Federal Reserve's announcement in October 2019 that it would maintain a level of reserve supply greater than the one that prevailed in early September is consistent with the implications of our framework.

In monetary policymaking, central bankers have long pointed out the importance of measuring the expectations of financial market participants, households, and firms especially with regard to inflation and the central bank's so-called "reaction function" to changes in the economic outlook. In addition to model-and market-implied measures, there has been a growing interest in and reliance on survey-based measures of subjective expectations. This article describes two major innovative survey initiatives conducted by the New York Fed to measure policy-relevant expectations of households and market participants: the Survey of Consumer Expectations, and the Survey of Primary Dealers and Survey of Market Participants. A key feature of these surveys is its use of a probabilistic question format to elicit the likelihood respondents assign to different future events. We discuss the advantages of using probabilistic questions, illustrate their value in more fully measuring beliefs and uncertainty, and document the pervasiveness and importance of heterogeneity in beliefs among our survey respondents.

Keynote remarks for the Commemoration of the Centennial of the Federal Reserve?s U.S. Dollar Account Services to the Global Official Sector, Federal Reserve Bank of New York, New York City.

A measure of underlying inflation that uses all relevant information, is available in real time, and forecasts inflation better than traditional underlying inflation measures?such as core inflation measures?would greatly benefit monetary policymakers, market participants, and the public. This article presents the New York Fed Staff Underlying Inflation Gauge (UIG) for the consumer price index and the personal consumption expenditures deflator. Using a dynamic factor model approach, the UIG is derived from a broad data set that extends beyond price series to include a wide range of nominal, real, and financial variables. This modeling approach also makes it possible to combine information simultaneously from the cross-sectional and time dimensions of the sample in a unified framework. In addition, the UIG can be updated on a daily basis to closely monitor changes in underlying inflation?a feature that is especially useful when sudden and large economic fluctuations occur, as was the case during the 2008 global financial crisis. Lastly, the UIG displays greater forecast accuracy than many measures of core inflation. Editor?s note: This article?s data appendix has been updated to reflect the removal of a duplicate price series (CPI-U: Other fresh vegetables). The article?s conclusions remain the same. (December 2017)

In this paper, we develop a bivariate unobserved components model for inflation and unemployment. The unobserved components are trend inflation and the non-accelerating inflation rate of unemployment (NAIRU). Our model also incorporates a time-varying Phillips curve and time-varying inflation persistence. What sets this paper apart from the existing literature is that we do not use unbounded random walks for the unobserved components, but rather use bounded random walks. For instance, trend inflation is assumed to evolve within bounds. Our empirical work shows the importance of bounding. We find that our bounded bivariate model forecasts better than many alternatives, including a version of our model with unbounded unobserved components. Our model also yields sensible estimates of trend inflation, NAIRU, inflation persistence and the slope of the Phillips curve.

Monetary policymakers and long-term investors would benefit greatly from a measure of underlying inflation that uses all relevant information, is available in real-time, and forecasts inflation better than traditional underlying inflation measures such as core inflation measures. This paper presents the Federal Reserve Bank of New York (FRBNY) Staff Underlying Inflation Gauge (UIG) for CPI and PCE. Using a dynamic factor model approach, the UIG is derived from a broad data set that extends beyond price series to include a wide range of nominal, real, and financial variables. It also considers the specific and time-varying persistence of individual subcomponents of an inflation series. An attractive feature of the UIG is that it can be updated on a daily basis, which allows for a close monitoring of changes in underlying inflation. This capability can be very useful when large and sudden economic fluctuations occur, as at the end of 2008. In addition, the UIG displays greater forecast accuracy than traditional measures of core inflation.

This paper uses multilevel factor models to characterize within- and between-block variations as well as idiosyncratic noise in large dynamic panels. Block-level shocks are distinguished from genuinely common shocks, and the estimated block-level factors are easy to interpret. The framework achieves dimension reduction and yet explicitly allows for heterogeneity between blocks. The model is estimated using an MCMC algorithm that takes into account the hierarchical structure of the factors. The importance of block-level variations is illustrated in a four-level model estimated on a panel of 445 series related to different categories of real activity in the United States.

To conduct monetary policy, central banks around the world increasingly rely on measures of public inflation expectations. In this article, we review findings from an ongoing initiative at the Federal Reserve Bank of New York aimed at improving the measurement and our understanding of household inflation expectations through surveys. We discuss the importance of question wording and the usefulness of new questions to elicit an individual’s distribution of inflation beliefs. We present evidence suggesting that consumers update their inflation expectations in response to new information and that information dissemination may lead to more informed and reliable reporting of inflation expectations. Finally, we report on a financially incentivized experiment suggesting that expectations surveys are informative and that respondents generally act on their stated beliefs in a way consistent with expected utility theory.

This chapter surveys the recent literature on output forecasting, and examines the real-time forecasting ability of several models for U.S. output growth. In particular, it evaluates the accuracy of short-term forecasts of linear and nonlinear structural and reduced-form models, and judgmental forecasts of output growth. Our emphasis is on using solely the information that was available at the time the forecast was being made, in order to reproduce the forecasting problem facing forecasters in real-time. We find that there is a large difference in forecast performance across business cycle phases. In particular, it is much harder to forecast output growth during recessions than during expansions. Simple linear and nonlinear autoregressive models have the best accuracy in forecasting output growth during expansions, although the dynamic stochastic general equilibrium model and the vector autoregressive model with financial variables do relatively well. On the other hand, we find that most models do poorly in forecasting output growth during recessions. The autoregressive model based on the nonlinear dynamic factor model that takes into account asymmetries between expansions and recessions displays the best real time forecast accuracy during recessions. Even though the Blue Chip forecasts are comparable, the dynamic factor Markov switching model has better accuracy, particularly with respect to the timing and depth of output fall during recessions in real time. The results suggest that there are large gains in considering separate forecasting models for normal times and models especially designed for periods of abrupt changes, such as during recessions and financial crises.

This article introduces a new model of trend inflation. In contrast to many earlier approaches, which allow for trend inflation to evolve according to a random walk, ours is a bounded model which ensures that trend inflation is constrained to lie in an interval. The bounds of this interval can either be fixed or estimated from the data. Our model also allows for a time-varying degree of persistence in the transitory component of inflation. In an empirical exercise with CPI inflation, we find the model to work well, yielding more sensible measures of trend inflation and forecasting better than popular alternatives such as the unobserved components stochastic volatility model. This article has supplementary materials online.

This paper introduces a new model of trend (or underlying) in‡ation. In contrast to many earlier approaches, which allow for trend in‡ation to evolve according to a random walk, ours is a bounded model which ensures that trend in‡ation is constrained to lie in an interval. The bounds of this interval can either be …xed or estimated from the data. Our model also allows for a time-varying degree of persistence in the transitory component of in‡ation. The bounds placed on trend in‡ation mean that standard econometric methods for estimating linear Gaussian state space models cannot be used and we develop a posterior simulation algorithm for estimating the bounded trend in‡ation model. In an empirical exercise with CPI in‡ation we …nd the model to work well, yielding more sensible measures of trend in‡ation and forecasting better than popular alternatives such as the unobserved components stochastic volatility model.

This appendix develops a posterior simulation algorithm for AR-trendbound: the bounded in‡ation model given in (5). The other models are restricted special cases of this model and, thus, the MCMC algorithm is restricted in the obvious manner in each case. The one exception of this is the UC-SV model of Stock and Watson (2007), which we label Trend-SV in the paper. This involves one extra state equation for the stochastic volatility in the in‡ation de ning trend in‡ation. This is drawn using the stochastic volatility described in this appendix. Except for the parameters a and b, the prior is described in Section 3.2. The priors for a and b are assumed to be uniform on the intervals (a; a) and (b; b) respectively, where a = 0, a = 1:5, b = 3:5 and b = 5. The MCMC algorithm sequentially draw from (we suppress the dependence on y0):

The relationship between short term and long term inflation expectations in the US and the UK is investigated with a focus on inflation pass through (i.e. how changes in short term expectations affect long term expectations). An econometric methodology is used which allows for the uncovering of the relationship between inflation pass through and various explanatory variables. Empirical results are related to theoretical models of anchored, contained and unmoored inflation expectations. For neither country are anchored or unmoored inflation expectations found. For the US, contained inflation expectations are found. For the UK, empirical findings are not consistent with the specific model of contained inflation expectations presented here, but are consistent with a broader view of expectations being constrained by the existence of an inflation target.

This paper develops a structured dynamic factor model for the spreads between London Interbank Offered Rate (LIBOR) and overnight index swap (OIS) rates for a panel of banks. Our model involves latent factors which reflect liquidity and credit risk. Our empirical results show that surges in the short term LIBOR-OIS spreads during the 2007–2009 financial crisis were largely driven by liquidity risk. However, credit risk played a more significant role in the longer term (twelve-month) LIBOR-OIS spread. The liquidity risk factors are more volatile than the credit risk factor. Most of the familiar events in the financial crisis are linked more to movements in liquidity risk than credit risk.

In many applications involving time-varying parameter VARs, it is desirable to restrict the VAR coefficients at each point in time to be non-explosive. This is an example of a problem where inequality restrictions are imposed on states in a state space model. In this paper, we describe how existing MCMC algorithms for imposing such inequality restrictions can work poorly (or not at all) and suggest alternative algorithms which exhibit better performance. Furthermore, we show that previous algorithms involve an approximation relating to a key prior integrating constant. Our algorithms are exact, not involving this approximation. In an application involving a commonly used U.S. data set, we present evidence that the algorithms proposed in this paper work well.

This paper uses multi-level factor models to characterize withinand between-block variations as well as idiosyncratic noise in large dynamic panels. Block-level shocks are distinguished from genuinely common shocks, and the estimated block-level factors are easy to interpret. The framework achieves dimension reduction and yet explicitly allows for heterogeneity between blocks. The model is estimated using an MCMC algorithm that takes into account the hierarchical structure of the factors. A four-level model is estimated to study blockand aggregate-level dynamics in a panel of 445 series related to different categories of real activity in the United States. The model illustrates the importance of block-level variations in the data.

This technical appendix to the discussion paper provides a detailed description of the Gibbs sampler for Bayesian estimation, the draw parameters for the measurement and latent risk equations, as well as priors and prior sensitivity analysis.