nep-mon New Economics Papers
on Monetary Economics
Issue of 2016‒09‒04
thirty-one papers chosen by
Bernd Hayo
Philipps-Universität Marburg

  1. Dynamics of Overnight Money Markets : What Has Changed at the Zero Lower Bound? By Elizabeth C. Klee; Zeynep Senyuz; Emre Yoldas
  2. Credibility of Optimal Forward Guidance at the Interest Rate Lower Bound By Taisuke Nakata
  3. Managing Heterogeneous and Unanchored Expectations: A Monetary Policy Analysis By Hommes, C.H.; Lustenhouwer, J.
  4. Negative interest rate policies: Sources and implications By Carlos Arteta; M. Ayhan Kose; Marc Stocker; Temel Taskin
  5. Where do I see the Monetary Policy Normalization Tools on the Fed's Balance Sheet and Income Statement? By Christian Miller; Casey Clark
  6. Warehouse Banking By Giorgia Piacentino; Anjan Thakor; Jason Donaldson
  7. The Expected Real Interest Rate in the Long Run : Time Series Evidence with the Effective Lower Bound By Benjamin K. Johannsen; Elmar Mertens
  8. Qualitative Guidance and Predictability of Monetary Policy in South Africa By Alain Kabundi; NtuthukoTsokodibane
  9. Risk shocks close to the zero lower bound By Seneca, Martin
  10. Hanging on Every Word : Semantic Analysis of the FOMC's Postmeeting Statement By Ellen E. Meade; Miguel Acosta
  11. Monetary versus macroprudential policies causal impacts of interest rates and credit controls in the era of the UK Radcliffe Report By Aikman, David; Bush, Oliver; Davis, Alan
  12. The inherent benefit of monetary unions By Groll, Dominik; Monacelli, Tommaso
  13. Inflation at the Household Level By Sam Schulhofer-Wohl; Greg Kaplan
  14. Sustainable International Monetary Policy Cooperation By Ippei Fujiwara; Timothy Kam; Takeki Sunakawa
  15. The Effects of FOMC Communications before Policy Tightening in 1994 and 2004 By Ellen E. Meade; Yoshio Nozawa; Lubomir Petrasek; Joyce Zickler
  16. The Federal Reserve's Overnight and Term Reverse Repurchase Agreement Operations in the Financial Accounts of the United States By Ralf R. Meisenzahl
  17. Did the Fed's Announcement of an Inflation Objective Influence Expectations? By Alan K. Detmeister; Daeus Jorento; Emily Massaro; Ekaterina V. Peneva
  18. Leaning Within a Flexible Inflation-Targeting Framework: Review of Costs and Benefits By Denis Gorea; Oleksiy Kryvtsov; Tamon Takamura
  19. Low long-term interest rates as a global phenomenon By Peter Hördahl; Jhuvesh Sobrun; Philip Turner
  20. The Currency Composition of Sovereign Debt By Diego Perez; Pablo Ottonello
  21. Monetary Policy and Asset Valuation: Evidence From a Markov-Switching cay By Francesco Bianchi; Martin Lettau; Sydney C. Ludvigson
  22. The Effects of Liquidity Regulation on Participation in the Term Deposit Facility By Marcelo Rezende; Mary-Frances Styczynski
  23. Reflections on the natural rate of interest, its measurement, monetary policy and the zero lower bound By Cukierman, Alex
  24. Toward Removal of the Swiss Franc Cap: Market Expectations and Verbal Interventions By Nikola Mirkov; Igor Pozdeev; Paul Söderlind
  25. The FOMC Meeting Minutes : An Assessment of Counting Words and the Diversity of Views By Ellen E. Meade; Nicholas A. Burk; Melanie Josselyn
  26. Assessing the effects of unconventional monetary policy on pension funds risk incentives By Boubaker, Sabri; Gounopoulos, Dimitrios; Nguyen, Duc Khuong; Paltalidis, Nikos
  27. The Risk of Returning to the Effective Lower Bound : An Implication for Inflation Dynamics after Lift-Off By Timothy S. Hills; Taisuke Nakata; Sebastian Schmidt
  28. Price plans and the real effects of monetary policy By Fernando Alvarez; Francesco Lippi
  29. Monetary Policy, Investment, and Firm Heterogeneity By Thomas Winberry; Pablo Ottonello
  30. Causal Change Detection in Possibly Integrated Systems: Revisiting the Money-Income Relationship By Shuping Shi; Stan Hurn; Peter C B Phillips
  31. Step away from the zero lower bound: Small open economies in a world of secular stagnation By Corsetti, G.; Mavroeidi, E.; Thwaites, G.; Wolf, M.

  1. By: Elizabeth C. Klee; Zeynep Senyuz; Emre Yoldas
    Abstract: Print{{p}}December 21, 2015{{p}}Dynamics of Overnight Money Markets: What Has Changed at the Zero Lower Bound?*{{p}}Elizabeth Klee, Zeynep Senyuz, and Emre Yoldas{{p}}The primary instrument of monetary policy in the U.S. has been the interest rate at which depository institutions (DIs) lend balances to{{p}}each other overnight, the federal funds rate (FFR). Prior to the 2007-2009 financial crisis, daily open market operations had been{{p}}conducted to keep the equilibrium FFR near the target rate determined by the Federal Open Market Committee (FOMC). The response{{p}}of the Federal Reserve to the crisis significantly changed the landscape for implementation of monetary policy. The level of reserves in{{p}}the banking system reached unprecedented levels as a variety of new facilities and tools were used to provide liquidity to the financial{{p}}system.1{{p}}Against this backdrop, the Federal Reserve started paying interest on excess reserves (IOER) held by DIs at the end of 2008. Money{{p}}market rates have been near the effective zero lower bound (ZLB) since the target range of 0 to 0.25 percent for the FFR was{{p}}established in December 2008. In September 2014, the FOMC indicated that during the normalization of the stance of monetary policy, it{{p}}intends to move the FFR into a target range mainly by adjusting the IOER and use an overnight reverse repurchase (ON RRP) facility{{p}}and other supplementary tools as needed.2{{p}}In this note we provide a comparative analysis of overnight money market dynamics before the crisis and after the target FFR has been{{p}}lowered to the ZLB. In addition, we also zoom into the ZLB period and analyze the two sub-periods before and after the ON RRP facility{{p}}test operations have started in September 23, 2013. Overall, we find that the FFR has continued to be interconnected with the other{{p}}money market rates although the co-movement seems to have declined somewhat. The most notable changes in rate dynamics are{{p}}observed on calendar days including the financial reporting dates as well as the days of the reserve maintenance period of DIs.3 We{{p}}show that volatility in the repo market has substantially declined since the inception of the ON RRP facility.{{p}}Data and Methodology{{p}}The data set consists of five overnight money market rates, namely the FFR, the London Interbank Offered Rate (Libor), the Eurodollar{{p}}rate (ED), the primary dealer survey repo rate (RP), and the interest rate on AA-rated nonfinancial commercial paper (CP). We use the{{p}}daily series from January 2, 2001 to August 31, 2015, and exclude the part of the financial crisis episode until the beginning of the ZLB{{p}}period in December 2008.4 We use vector autoregressions and time-varying volatility models to analyze joint dynamics of the{{p}}aforementioned interest rates. In particular, the model for the pre-crisis period is a vector error correction (VEC) model that incorporates{{p}}the long-run equilibrium relationship of the rates as well as changes in the target FFR. For the ZLB period, given the stationary behavior{{p}}of rates, we estimate a vector autoregressive (VAR) specification in levels of the interest rates. Both models incorporate certain calendar{{p}}dates as well as the days of the maintenance period that may influence overnight funding rates. The ZLB model does not include target{{p}}FFR since it is constant for the entire period. In both cases, GARCH models with calendar effects are applied to model residuals to{{p}}capture time-varying volatility.{{p}}Results{{p}}Our estimates for the pre-crisis period are consistent with the monetary policy implementation framework in which FFR had been kept{{p}}near the FOMC target through open market operations, and market activity led to adjustments in other rates consistent with movements{{p}}in the FFR.5 In particular, we find that changes in the target FFR are highly significant in all rate equations, reflecting pass-through from{{p}}the policy rate to other overnight rates (Table 1, Panel A). Moreover, the parameters that quantify the adjustment of all interest rates{{p}}toward long-run equilibrium with the FFR are also highly significant, implying that it has been the other rates, not the FFR, that adjusted{{p}}in response to market fluctuations. We document significant autocorrelation dynamics in all rates, while cross lags are mostly{{p}}insignificant when we control for changes in target FFR and incorporate long-run equilibrium dynamics (Table 1, Panel B). At the ZLB,{{p}}although the FFR continued to provide an anchor for the unsecured overnight rates, the transmission to the repo rate is hampered as{{p}}implied by the insignificance of the lagged FFR in the RP equation (Table 2).{{p}}Table 1: The Pre-Crisis Model for Overnight Money Market Rates{{p}}Panel A. FFR Target Change and Error Correction Terms{{p}}FFR RP Libor CP{{p}}D(FFR) 0.454 0.406 0.337 0.416{{p}}(0.00) (0.00) (0.00) (0.00){{p}} FRB: FEDS Notes: Dynamics of Overnight Money Markets: What Has ...{{p}}1 of 5 12/28/2015 8:08 AM{{p}}FFR RP Libor CP{{p}}EC1 -0.032 0.303 -0.01 -0.001{{p}}(0.39) (0.00) (0.72) (0.98){{p}}EC2 -0.016 0.125 0.455 0.091{{p}}(0.91) (0.41) (0.00) (0.43){{p}}EC3 -0.005 0.021 -0.1 0.431{{p}}(0.97) (0.92) (0.21) (0.01){{p}}Panel B. Sum of Autoregressive Terms{{p}}FFR RP Libor CP{{p}}FFR -0.396 -0.057 -0.266 -0.12{{p}}(0.03) (0.82) (0.06) (0.58){{p}}RP -0.041 -0.692 0.171 0.071{{p}}(0.65) (0.00) (0.02) (0.42){{p}}Libor 0.728 0.258 -0.749 0.281{{p}}(0.04) (0.48) (0.00) (0.22){{p}}CP -0.767 0.143 0.506 -0.65{{p}}(0.06) (0.79) (0.02) (0.07){{p}}Estimates from a VEC model are reported. The daily sample runs from January 2, 2001 to July 31, 2007. P-values based on robust (HAC) standard errors are{{p}}reported in parenthesis. Lag length selected by Schwarz information criterion is 4. EC1, EC2, and EC3 denote the error correction terms implied by the{{p}}long-run equilibrium relationship of FFR with RP, Libor, and CP respectively.{{p}}Table 2: The ZLB Model for Overnight Money Market Rates{{p}}Sum of Autoregressive Terms{{p}}FFR RP Libor CP{{p}}FFR 0.911 0.107 0.223 0.153{{p}}(0.00) (0.21) (0.00) (0.02){{p}}RP 0.032 0.809 0.014 -0.011{{p}}(0.00) (0.00) (0.29) (0.47){{p}}ED -0.024 0.048 0.705 0{{p}}(0.53) (0.50) (0.00) (1.00){{p}}CP 0.036 0.054 0.069 0.881{{p}}(0.01) (0.14) (0.00) (0.00){{p}}Estimates from a VAR model are reported. The daily sample runs from December 17, 2008 to August 31, 2015. p-values based on robust (HAC) standard{{p}}errors are reported in parenthesis. Lag length selected by Schwarz criterion is 4.{{p}}The correlations of rates that are estimated after accounting for persistence and volatility dynamics also suggest continued{{p}}interconnectedness of FFR to other interest rates at the ZLB (Figure 1). The correlations of FFR with RP and CP declined somewhat at{{p}}the ZLB period, but remained sizable. However, financial reporting dates appear to have substantial effects on the comovement of FFR{{p}}and RP against the backdrop of abundant bank reserves and changing financial regulations. For example, the correlation of FFR and RP{{p}}on quarter-ends declined from 0.35 to effectively 0 at the ZLB.6{{p}}Figure 1: Correlations of FFR with the Other Rates: Pre-crisis vs. ZLB{{p}} FRB: FEDS Notes: Dynamics of Overnight Money Markets: What Has ...{{p}}2 of 5 12/28/2015 8:08 AM{{p}}Dots indicate point estimates and the surrounding bands are 95% confidence intervals. N denotes normal days that exclude month-end and quarter-end dates.{{p}}M and Q denote month-end and quarter-end respectively. Pre-crisis sample includes Libor, which is replaced with ED in the ZLB period.{{p}}Accessible version{{p}}Another notable change in the ZLB period relative to the pre-crisis era has been the disappearance of the day-of-maintenance-period{{p}}effects on the FFR, likely reflecting the abundance of bank reserves in the system. For example, the FFR used to be firmer on Mondays{{p}}possibly due to elevated payment flows while softer on Fridays since banks usually try to avoid an excess position over the weekend{{p}}during which reserves count for three days of requirement.7 In contrast, we find no statistically or economically significant day-of-maintenance-{{p}}period effects on the FFR in the ZLB period.{{p}}The dynamics surrounding financial reporting days changed notably at the ZLB amid abundant reserves in the banking system as well{{p}}as the introduction of new financial regulations. All rates were subject to modest upward pressure at month-ends prior to the crisis while{{p}}quarter-end effects had been somewhat more prominent (Figure 2, left). The most notable quarter-end effect had been the decline in the{{p}}repo rate, which likely reflected window-dressing activity that decreased demand for repo financing on such dates. The unsecured{{p}}financing rates had exhibited relatively smaller movements on quarter-ends, and in the opposite direction. In contrast, the quarter-end{{p}}effects turned negative for all unsecured rates at the ZLB amid the announcement and implementation of Basel III capital and liquidity{{p}}reforms. In particular, the Liquidity Coverage Ratio and leverage requirements seem to have reduced materially banks' demand for{{p}}short-term unsecured borrowing on reporting days.{{p}}Figure 2: Calendar Effects on Overnight Money Market Rates: Pre-crisis vs. ZLB{{p}}Dots indicate point estimates and the surrounding bands are 95% confidence intervals. M and Q denote month-end and quarter-end respectively. Effects are{{p}}normalized with respect to the standard deviations of model residuals.{{p}}Accessible version{{p}}Figure 3: RP Volatility and ON RRP{{p}} FRB: FEDS Notes: Dynamics of Overnight Money Markets: What Has ...{{p}}3 of 5 12/28/2015 8:08 AM{{p}}Estimated volatility of the RP from a GARCH model with month-end and quarter-end effects.{{p}}Accessible version{{p}}Contrary to the case of unsecured rates, the quarter-end effect has become insignificant for RP in the ZLB period. This estimate{{p}}captures the net effect as the ZLB period contains several quarter-end dates with material movements in RP in either direction. Such{{p}}changes are reflected in the statistically significant quarter-end effect on the volatility of the repo rate during the ZLB, which we estimate{{p}}to be around 2 basis points. As shown in Figure 3 the decline in RP volatility is strikingly evident both in the level and the volatility of the{{p}}series. Consistent with the intended effect of ON RRP to set a soft floor for repo rates, volatility in the repo market has substantially{{p}}dampened after the introduction of the program. We also find that calendar effects on RP volatility largely disappeared after that point.{{p}}Again, this likely reflects the soft floor placed on repo rates by the ON RRP, which lessen the potential for sharp falls in rates, as well as{{p}}the availability of the ON RRP as a viable investment, especially on financial reporting dates when other options may not be available.{{p}}Both results point out to a significant change in the structure of the money markets initiated by the test operations of the ON RRP facility.{{p}}References{{p}}Bech, M., E. Klee, and V. Stebunovs (2014): "Arbitrage, Liquidity and Exit: The Repo and Federal Funds Market before, during and{{p}}Emerging from the Financial Crisis', in Developments in Macro-Finance Yield Curve Modelling, ed. by J.S. Chadha, A.C. J. Durr, M.A. S.{{p}}Joyce, and L. Sarno, Cambridge University Press, 293-325.{{p}}Carpenter, S. and S. Demiralp (2006): "The Liquidity Effect in the Federal Funds Market," Journal of Money, Credit and Banking, 38,{{p}}901-920.{{p}}Hamilton, J. D., (1996): "The Daily Market for Federal Funds," Journal of Political Economy 104, 26-56.{{p}}Judson, R. and E. Klee (2010): "Whither the Liquidity Effect: The Impact of Federal Reserve Open Market Operations in Recent Years,"{{p}}Journal of Macroeconomics, 32, 713-731.{{p}}* We thank James Clouse, Jane Ihrig, and Josh Louria for helpful comments and Richard Sambasivam for research assistance. Return to text{{p}}1. The reserve balances of depository institutions currently stand close to $3 trillion compared to the pre-crisis level of about $25 billion. Return to text{{p}}2. See "Policy Normalization Principles and Plans" issued by the Federal Reserve on September 17, 2014{{p}}/monetary/20140917c.htm. Return to text{{p}}3. An institution is responsible for satisfying its reserve balance requirement by holding balances on average over a 14-day maintenance period in an account{{p}}at the Federal Reserve. For details of reserve maintenance, see Return to text{{p}}4. The daily effective FFR is calculated as the volume-weighted average of rates on trades arranged by major brokers and is available from the FRBNY. The{{p}}Treasury general collateral repo rate is a volume-weighted average on overnight repo transactions where the underlying collateral is U.S. Treasury security{{p}}and it is obtained from the primary dealer survey of the FRBNY. We use the ED data that the FRBNY started to publish in March 2010. Prior to that date, the{{p}}ED data are obtained from ICAP. For the pre-crisis period, Libor is substituted for the ED since the latter is not available. The CP rate is from the Federal{{p}}Reserve Board's CP release which is derived from data supplied by The Depository Trust and Clearing Corporation (DTCC). Return to text{{p}}5. See Bech et al. (2014) for a detailed analysis of repo and federal funds markets. Return to text{{p}}6. We also observe substantially lower correlations of the secured RP with other unsecured rates on financial reporting dates at the ZLB (not shown). Return to{{p}}text{{p}}7. See for example, Hamilton (1996), Carpenter and Demiralp (2006) and, Judson and Klee (2010). Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}} FRB: FEDS Notes: Dynamics of Overnight Money Markets: What Has ...{{p}}4 of 5 12/28/2015 8:08 AM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}Last update: December 21, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: Dynamics of Overnight Money Markets: What Has ...{{p}}5 of 5 12/28/2015 8:08 AM
    Date: 2015–12–21
  2. By: Taisuke Nakata
    Abstract: Figure 1: The Value of Commitment in a Stylized Model{{p}}August 27, 2015{{p}}Credibility of Optimal Forward Guidance at the Interest Rate Lower Bound{{p}}Taisuke Nakata{{p}}1. Introduction{{p}}Market participants and other analysts generally expect that the federal funds rate will rise from its effective lower bound (ELB) later this{{p}}year.1 However, the ELB could again become a binding constraint on monetary policy in the future.2 The ELB constraint prevents central{{p}}banks from further stimulating the economy through conventional means, making the economic conditions worse than they would{{p}}otherwise be. Accordingly, developing an effective strategy to address the adverse consequences of the lower bound constraint remains{{p}}an important task for economists and policymakers.{{p}}In this note, I first describe the effectiveness of a particular form of forward guidance policy--optimal commitment policy--in mitigating the{{p}}adverse consequences of the ELB. I then describe a key criticism against adopting this policy in reality--namely, that the policy is{{p}}potentially not credible. Finally, I discuss my research, Nakata (2014), that investigates how the central bank's concern for reputation can{{p}}overcome the credibility problem of the commitment policy.{{p}}2. Optimal Commitment Policy: A Strong Form of Forward Guidance{{p}}In many macroeconomic models, optimal commitment policy is very effective in mitigating the adverse consequences of the lower bound{{p}}constraint. Under this policy, the central bank commits to keeping its policy rate at the ELB for an extended period, with the explicit goal{{p}}of temporarily leading inflation and the output gap to overshoot their longer-run targets. In economies where households and price-setters{{p}}are forward looking, the temporary overheating of the economy tempers the declines in inflation and output during the period{{p}}when the lower bound is a constraint through improved expectations. This policy can be thought of as a particularly strong form of{{p}}forward guidance policy and has been analyzed by many economists.3 In particular, Michael Woodford of Columbia University made the{{p}}case for optimal commitment policy as a practical policy tool in a Jackson Hole presentation in 2012.4{{p}}Figure 2{{p}}illustrates{{p}}how this{{p}}commitment{{p}}policy works{{p}}in the{{p}}stylized{{p}} FRB: FEDS Notes: Credibility of Optimal Forward Guidance at the Intere...{{p}}1 of 6 8/27/2015 11:45 AM{{p}}Accessible version{{p}}macroeconomic model I use in Nakata (2014). In these simulations, there is a one-period crisis shock that hits the economy at period{{p}}one and disappears at period two. The blue and red lines show the dynamics of the federal funds rate, inflation, and the output gap{{p}}under optimal discretionary policy and under optimal commitment policy, respectively.5 Under the discretionary policy, the policy rate is{{p}}kept at the lower bound while the crisis shock lasts but returns to the steady state as soon as the crisis shock disappears. Inflation and{{p}}the output gap decline by 2 percentage points and 10 percentage points at period one, respectively. At period two, inflation is back to{{p}}target and the output gap is zero.{{p}}Under the commitment policy, the central bank keeps the policy rate at the ELB until period four. This policy generates an overshooting{{p}}of inflation and output at period two and beyond. Because price-setting is forward looking in this model, the higher expected inflation{{p}}after the crisis leads to higher actual inflation in the period of the crisis through the expectations term in the Phillips curve of the model.{{p}}Similarly, because households are forward looking, the higher expected output gap and lower expected real rate associated with higher{{p}}inflation mitigate the decline in the output gap in the crisis period through the expectations terms in the aggregate demand equation.{{p}}Inflation and the output gap thus decline only 0.3 and 6 percentage points respectively at period one.{{p}}3. A Case against the Commitment Policy: Lack of Credibility{{p}}While the commitment policy is effective in mitigating the adverse effects of the lower bound constraint, there is an important caveat:{{p}}Commitment may not be credible. That is, when the central bank announces this policy at the onset of the crisis, the private sector may{{p}}not believe that the central bank will stick to its commitment in the future. This tension arises because the central bank will have an{{p}}incentive to renege on the commitment. While the central bank wants to promise an extended period of low policy rates at the onset of{{p}}the crisis, once the crisis is over, the central bank is better off raising the policy rate and eliminating the overshooting of inflation and{{p}} FRB: FEDS Notes: Credibility of Optimal Forward Guidance at the Intere...{{p}}2 of 6 8/27/2015 11:45 AM{{p}}Figure 2. Costs and Benefits of Reneging on the Promise{{p}}output because such overshooting is undesirable ex post. In the academic literature, when the government ex post has an incentive to{{p}}renege on its promised policy action and thus the private sector does not believe in the government's promise, the policy is said to be{{p}}time-inconsistent. The credibility problem just described is a particular example of time-inconsistency.6{{p}}This time-inconsistency problem of optimal commitment policy is not a mere theoretical curiosity. This problem has been cited by{{p}}policymakers as a factor limiting the effectiveness of stimulating the economy through announcements about future policy actions. For{{p}}example, John Williams, president of the Federal Reserve Bank of San Francisco, stated:{{p}}The optimal forward guidance policy is not time-consistent. According to the theory, for this policy to have the desired effects, the central{{p}}bank must commit to two things: keeping the short-term policy rate lower than it otherwise would in the future, and allowing inflation to{{p}}rise higher than it otherwise would. However, when the time comes for the central bank to fulfill this commitment, it may not want to do{{p}}so. It might find it hard to resist the temptation to raise rates earlier than promised to avoid the rise in inflation. [Williams (2012)]{{p}}James Bullard, president of the Federal Reserve Bank of St. Louis, also acknowledges the difficulty of credibly committing to keeping the{{p}}policy rate low for long and permitting inflation to overshoot:{{p}}The "Woodford period" approach to forward guidance [i.e., optimal commitment policy] relies on a credible announcement made today{{p}}that future monetary policy will deviate from normal. The central bank does not actually behave differently today. One might argue that{{p}}such an announcement is unlikely to be believed. Why should future monetary policy deviate from normal once the economy is growing{{p}}and inflation is rising? But if the announcement is not credible, then the private sector will not react with more consumption and{{p}}investment today. That is, any effects would be minimal. [Bullard (2013)]{{p}}Similarly, Mark Carney, former governor of the Bank of Canada and current governor of the Bank of England, stated:{{p}}Today, to achieve a better path for the economy over time, a central bank may need to commit credibly to maintaining highly{{p}}accommodative policy even after the economy and, potentially, inflation picks up. Market participants may doubt the willingness of an{{p}}inflation-targeting central bank to respect this commitment if inflation goes temporarily above target. These doubts reduce the effective{{p}}stimulus of the commitment and delay the recovery. [Carney (2012)]{{p}}Some policymakers see this time-inconsistency problem as a key factor that makes central banks reluctant to adopt this commitment{{p}}policy in practice. According to Benoît Coeuré, board member at the European Central Bank,{{p}}The main challenge of such guidance [i.e., optimal commitment policy] is its inherent inconsistency over time and thus lack of credibility.{{p}}... This is a possible explanation why, in practice, central banks have refrained from using forward guidance in a way that implies a major{{p}}change in strategy. Therefore, central banks' forward guidance has rather aimed at providing greater clarity on the reaction function and{{p}}the assessment of future economic conditions. [Coeuré (2013)]{{p}}These quotes are only a few of many instances in which central bank officials have expressed their concern about the credibility problem{{p}}of the commitment policy, suggesting that concerns about credibility are an important consideration for many policymakers.7{{p}}4. Reputation as a Way to Overcome the Credibility Problem{{p}}In my research, I study whether the central bank's concern for reputation can make the commitment policy credible. My model combines{{p}}a theory of reputation from the game theory literature with a standard sticky-price macroeconomic model. In my model, if the central{{p}}bank reneges on its promise to keep the policy rate low for an extended period, it can eliminate overshooting of inflation and output in{{p}}the short run but it loses its reputation and the private sector will not believe similar promises in future recessions. Instead, the private{{p}}sector believes that the central bank will follow the discretionary policy, and the central bank loses its ability to conduct the commitment{{p}}policy. As just described, the discretionary policy would entail worse outcomes for inflation and the output gap than the commitment{{p}}policy. Thus, a concern for reputation creates an incentive for the central bank to fulfill its promises.8{{p}}However,{{p}}certain{{p}}criteria must{{p}}be met for{{p}}the{{p}}reputational{{p}}mechanism{{p}}to work. One{{p}}key{{p}}condition is{{p}}that the{{p}}crisis shock{{p}} FRB: FEDS Notes: Credibility of Optimal Forward Guidance at the Intere...{{p}}3 of 6 8/27/2015 11:45 AM{{p}}Accessible version{{p}}hits the{{p}}economy{{p}}sufficiently{{p}}frequently.{{p}}To illustrate{{p}}this point,{{p}}Figure 2{{p}}shows how{{p}}the crisis{{p}}probability{{p}}affects the{{p}}short-run{{p}}benefit and{{p}}the long-run{{p}}cost of{{p}}reneging on{{p}}the promise{{p}}of keeping{{p}}the policy{{p}}rate at the{{p}}lower bound{{p}}in the period{{p}}after the{{p}}crisis shock{{p}}disappears.{{p}}Here, the{{p}}short-run{{p}}benefit of{{p}}reneging on{{p}}the promise, shown by the red line, is how much the central bank gains by eliminating the overshooting of inflation and the output gap.{{p}}This short-run benefit does not depend on the crisis probability, hence this line is flat. The long-run cost, shown by the black line, is the{{p}}cost to the central bank of losing its reputation and thus its ability to conduct the commitment policy in the future. Because the benefits of{{p}}the commitment policy accrue during crises, this long-run cost increases with the crisis probability. Thus, the commitment policy is{{p}}credible only if the crisis probability is sufficiently high. In the example presented, the threshold crisis probability above which the{{p}}commitment policy is credible is very small, less than 0.1 percent per quarter. In the U.S., a naïve estimate of this crisis probability over{{p}}the past one hundred years would be about 0.5 percent per quarter, as there were two large shocks (the Great Depression and the{{p}}Great Recession) in which the Federal Reserve lowered the federal funds rate to its ELB in the 100 years since the creation of the{{p}}Federal Reserve (0.005=2/(4*100)).{{p}}I find that even a very small crisis probability is enough to make the commitment policy credible under many alternative assumptions{{p}}about the structure of the economy. This is true even when I modify the model to allow the chair at the central bank to change over time{{p}}so that loss of reputation is temporary. Thus, taken at face value, my results suggest that, if a central bank were to engage in this{{p}}commitment policy, reputational forces are likely to be strong enough to make it credible.{{p}}5. Conclusion{{p}}Optimal commitment policy has recently attracted a lot of attention among economists and policymakers as a potentially effective{{p}}approach to stimulating the economy when the short-term policy rate is constrained at the ELB. Despite its effectiveness in many{{p}}macroeconomic models, a number of central bank officials have expressed reservations about adopting such policies. One key reason{{p}}policymakers have reservations is this policy's time-inconsistency. My results suggest that a central bank's concern for its reputation can{{p}}make commitment policy time-consistent, potentially alleviating such concerns.{{p}}My framework permits the analysis of one key issue policymakers may want to consider for adopting the optimal commitment policy--{{p}}credibility. However, the validity of various assumptions responsible for making this optimal commitment policy effective would need to{{p}}be examined carefully if a central bank were to adopt this policy. For example, one key assumption of this theory is that households and{{p}}price-setters are forward looking and have rational expectations. Another is that the private sector correctly understands that the{{p}}overshooting of inflation from its longer-run target is temporary, and that long-run inflation expectations remain anchored at the central{{p}}bank's target.{{p}}Finally, while I have focused on the credibility of optimal commitment policy, my analysis is also useful in thinking about the credibility of{{p}}certain policy rules that share key aspects of the commitment policy, such as price-level targeting and nominal-income targeting rules.{{p}}Like the commitment policy, these rules also imply that the policy rate is kept at the lower bound for an extended period and that inflation{{p}}and the output gap overshoot their targets. My research suggests that, like the optimal commitment policy, a concern for reputation{{p}} FRB: FEDS Notes: Credibility of Optimal Forward Guidance at the Intere...{{p}}4 of 6 8/27/2015 11:45 AM{{p}}would make these rules credible should a central bank choose to adopt them.{{p}}References{{p}}Adam, K., and R. Billi (2006): "Optimal Monetary Policy Under Commitment with a Zero Bound on Nominal Interest Rates," Journal of{{p}}Money, Credit, and Banking, 38(7), 1877–1905.{{p}}Ball, L. M. (2013): "The Case for Four Percent Inflation," Central Bank Review, 13, 17–31.{{p}}Barro, R., and D. Gordon (1983): "Rules, Discretion, and Reputation in a Model of Monetary Policy," Journal of Monetary Economics, 12,{{p}}101–121.{{p}}Bean, C. (2013): "Global Aspects of Unconventional Monetary Policies," Remarks at the Federal Reserve Bank of Kansas City{{p}}Economic Policy Symposium, Jackson Hole, Wyoming.{{p}}Bullard, J. (2013): "Monetary Policy in a Low Policy Rate Environment," OMFIF Golden Series Lecture, London, United Kingdom.{{p}}Carney, M. (2012): "Guidance," Remarks at the CFA Society Toronto, Toronto, Ontario.{{p}}Coeuré, B. (2013): "The Usefulness of Forward Guidance," Remarks at the Money Marketeers Club of New York, New York City, New{{p}}York.{{p}}Dudley, W. (2013): "Remarks at the Central Bank Independence Conference: Progress and Challenges in Mexico," Remarks at the{{p}}Central Bank Independence Conference: Progress and Challenges in Mexico, Mexico City, Mexico.{{p}}Eggertsson, G., and M. Woodford (2003): "The Zero Bound on Interest Rates and Optimal Monetary Policy," Brookings Papers on{{p}}Economic Activity, 34(1), 139–235.{{p}}Jung, T., Y. Teranishi, and T. Watanabe (2005): "Optimal Monetary Policy at the Zero-Interest- Rate Bound," Journal of Money, Credit,{{p}}and Banking, 35(7), 813–35.{{p}}Kydland, F., and E. C. Prescott (1977): "Rules Rather than Discretion: The Inconsistency of Optimal Plans," Journal of Political Economy,{{p}}85(3), 473–493.{{p}}Lacker, J. (2013): "Monetary Policy in the United States: The Risks Associated With Unconventional Policies," Remarks at the{{p}}Swedbank Economic Outlook Seminar, Stockholm, Sweden.{{p}}Nakata, T. (2014): "Reputation and Liquidity Traps," Finance and Economics Discussion Series 2014-50, Board of Governors of the{{p}}Federal Reserve System (U.S.).{{p}}Nakov, A. (2008): "Optimal and Simple Monetary Policy Rules with Zero Floor on the Nominal Interest Rate," International Journal of{{p}}Central Banking, 4(2), 73–127.{{p}}Plosser, C. (2013): "Forward Guidance," Remarks at Stanford Institute for Economic Policy Re- searchs (SIEPR) Associates Meeting,{{p}}Stanford, California.{{p}}Rogoff, K. (1987): "Reputational Constraints on Monetary Policy," Carnegie-Rochester Conference Series on Public Policy, 26, 141–182.{{p}}Werning, I. (2012): "Managing a Liquidity Trap: Monetary and Fiscal Policy," Working Paper.{{p}}Williams, J. C. (2011): "Unconventional Monetary Policy: Lessons from the Past Three Years," FRBSF Economic Letter, 2011-31{{p}}(October 3).{{p}} ----(2012): "The Federal Reserve's Unconventional Policies," FRBSF Economic Letter, 2012-34 (November 13).{{p}}Woodford, M. (2012): "Methods of Policy Accommodation at the Interest-Rate Lower Bound," Presented at the 2012 Jackson Hole{{p}}Symposium, Federal Reserve Bank of Kansas City, Jackson Hole, Wyoming.{{p}}* I would like to thank Stephanie Aaronson, Eric Engen, David Lebow, Matthias Paustian, John Roberts, Gisela Rua, Robert Tetlow, and Bill Wascher for their{{p}} FRB: FEDS Notes: Credibility of Optimal Forward Guidance at the Intere...{{p}}5 of 6 8/27/2015 11:45 AM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}thoughtful comments. Timothy Hills and Paul Yoo provided excellent research assistance. The views expressed in this note, and all errors and omissions,{{p}}should be regarded as those solely of the author, and do not necessarily reflect those of the Federal Reserve Board of Governors or the Federal Reserve{{p}}System.{{p}}** Division of Research and Statistics, Board of Governors of the Federal Reserve System, 20th Street and Constitution Avenue N.W., Washington, D.C.{{p}}20551; Email:{{p}}1. In the 2015 June Survey of Primary Dealers, respondents on average put about 80 percent probability to the event that the federal funds rate will rise from{{p}}the ELB by the end of 2015. Return to text{{p}}2. For example, Ball (2013) argues that "the lower bound on interest rates is likely to constrain monetary policy in a large fraction of recessions'' in the United{{p}}States. Return to text{{p}}3. See Adam & Billi (2006); Eggertsson & Woodford (2003); Jung et al. (2005); Nakov (2008); and Werning (2012). Return to text{{p}}4. See Woodford (2012). Return to text{{p}}5. Under the discretionary policy, the central bank optimizes its strategy every period based on the economic conditions that prevail at that time. Under the{{p}}commitment policy, the central bank optimally designs its strategy at period one and commits to implementing that strategy afterward. Return to text{{p}}6. Time-inconsistency of optimal commitment policy was first noticed by Kydland & Prescott (1977). The problem of time-inconsistency arises in many other{{p}}contexts when private-sector agents are forward looking. Return to text{{p}}7. See Bean (2013), Dudley (2013), Lacker (2013), Plosser (2013), and Williams (2011) for other examples. Return to text{{p}}8. The idea of making commitment policies credible by introducing reputational forces has a long history. Most famously, Barro & Gordon (1983) and Rogoff{{p}}(1987) used the same idea to ask whether a central bank can credibly commit to low inflation in the model where the central bank has short-run incentives to{{p}}create surprise inflation. Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: August 27, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: Credibility of Optimal Forward Guidance at the Intere...{{p}}6 of 6 8/27/2015 11:45 AM
    Date: 2015–08–27
  3. By: Hommes, C.H. (University of Amsterdam); Lustenhouwer, J. (University of Amsterdam)
    Abstract: We study monetary policy in a New Keynesian model with heterogeneity in expectations. Agents may choose from a continuum of forecasting rules and adjust their expectations based on relative past performance. The extent to which expectations are anchored to the fundamentals of the economy turns out to be crucial in determining whether the central bank (CB) can stabilize the economy. When expectations are strongly anchored, little is required of the CB for local stability. Only when expectations are unanchored, the Taylor principle becomes a necessary condition. More aggressive policy may however be required to prevent coordination on almost self-fulfilling optimism or pessimism. When the zero lower bound on the nominal interest rate (ZLB) is accounted for, the inflation target must furthermore be high enough, in order to prevent coordination on self-fulfilling liquidity traps and deflationary spirals.
    Date: 2016
  4. By: Carlos Arteta; M. Ayhan Kose; Marc Stocker; Temel Taskin
    Abstract: Against the background of continued growth disappointments, depressed inflation expectations, and declining real equilibrium interest rates, a number of central banks have implemented negative interest rate policies (NIRP) to provide additional monetary policy stimulus over the past few years. This paper studies the sources and implications of NIRP. We report four main results. First, monetary transmission channels under NIRP are conceptually analogous to those under conventional monetary policy but NIRP present complications that could limit policy effectiveness. Second, since the introduction of NIRP, many of the key financial variables have evolved broadly as implied by the standard transmission channels. Third, NIRP could pose risks to financial stability, particularly if policy rates are substantially below zero or if NIRP are employed for a protracted period of time. Potential adverse consequences include the erosion of profitability of banks and other financial intermediaries, and excessive risk taking. However, there has so far been no significant evidence that financial stability has been compromised because of NIRP. Fourth, spillover implications of NIRP for emerging market and developing economies are mostly similar to those of other unconventional monetary policy measures. In sum, NIRP have a place in a policy maker’s toolkit but, given their domestic and global implications, these policies need to be handled with care to secure their benefits while mitigating risks.
    Keywords: Unconventional monetary policy, quantitative easing; bank profitability, financial stability, negative yields, event study, emerging markets, developing countries
    JEL: E52 E58 E60
    Date: 2016–08
  5. By: Christian Miller; Casey Clark
    Abstract: FEDS Notes Print{{p}}May 11, 2016{{p}}Where do I see the Monetary Policy Normalization Tools on the Fed's Balance Sheet and{{p}}Income Statement?{{p}}Christian Miller and Casey Clark1{{p}}On December 16, 2015, the Federal Open Market Committee (FOMC) determined it was appropriate to raise the effective federal funds{{p}}rate from the 0 to 25 basis point range it had been set at since late 2008. This note highlights where some of the key elements of the{{p}}FOMC's approach to policy normalization are reported on the Federal Reserve's website. Specifically, this note focuses on the interest on{{p}}excess reserves (IOER) rate, excess reserve balances, and interest expense on excess reserves. This note also identifies where{{p}}information can be found on the overnight reverse repurchase agreement (ON RRP) offering rate and the associated Federal Reserve{{p}}balances and interest expense.{{p}}Background on Federal Reserve publications{{p}}The Federal Reserve Board's Statistical Release H.4.1,"Factors Affecting Reserve Balances of Depository Institutions and Condition{{p}}Statement of Federal Reserve Banks," is a weekly publication that presents a balance sheet for each Federal Reserve Bank, a{{p}}consolidated balance sheet for all 12 Reserve Banks, an associated statement that lists the factors affecting reserve balances of{{p}}depository institutions, and several other tables presenting information on the assets, liabilities, and commitments of the Federal Reserve{{p}}Banks. Table 1 presents details on the factors that supply and absorb reserve balances, as well as the level of reserve balances--that is,{{p}}funds that depository institutions hold on deposit at the Federal Reserve to satisfy reserve requirements and funds held in excess of{{p}}requirements.2 Table 5 presents the balance sheet of the Federal Reserve System.{{p}}Annually, the Federal Reserve System releases the combined annual financial statements for the Federal Reserve Banks (combined{{p}}statements), as well as statements for the 12 individual Federal Reserve Banks, which provide a significant amount of information about{{p}}the assets, liabilities, earnings, and expenses of the Reserve Banks. The financial statements are audited annually by an independent{{p}}auditing firm. In addition to the annual financial statements, the Federal Reserve System releases quarterly the Federal Reserve Banks{{p}}Combined Quarterly Financial Reports, which provide unaudited quarterly updates to the information presented in the annual financial{{p}}statements.{{p}}IOER and reserve balances{{p}}The FOMC has stated that the IOER rate will be a primary tool during the normalization period.3 Depository institutions should be{{p}}unwilling to lend to any private counterparty at a rate lower than the rate they can earn on balances maintained at the Federal Reserve.{{p}}As a result, an increase in the IOER rate will put upward pressure on a range of short-term interest rates. In effect, raising the IOER rate{{p}}allows the Federal Reserve to increase the value that depository institutions place on reserve balances, which will have market effects{{p}}similar to those associated with a reduction in the quantity of reserves in the traditional, quantity-based mechanism for tightening the{{p}}stance of monetary policy.4 The IOER rate paid on excess reserve balances can be found on the Board of Governors' "Interest on{{p}}Required Balances and Excess Balances" page. Although the Federal Reserve pays interest on required reserves (IORR) in addition to{{p}}IOER, the marginal return of an additional dollar of reserves to a depository institution is the IOER rate given the large amount of excess{{p}}reserves in the System.{{p}}The H.4.1 reports the level of aggregate reserve balances--required and excess reserve balances together--on a weekly basis in Tables 1{{p}}and 5. Table 1 reports "Reserve balances with Federal Reserve Banks," which deducts depository institution overdrafts from overall{{p}}reserve balances.{{p}}Figure 1{{p}} FRB: FEDS Notes: Where do I see the Monetary Policy Normalization Tools on the Fed's... Page 1 of 5{{p}} 5/11/2016{{p}}In Table 5, the line item "Other deposits held by depository institutions" reports both required and excess reserve balances, but is not{{p}}reduced by any overdrafts which would be reported in the assets section of table 5. Thus, this value may not always align with "Reserve{{p}}balances with Federal Reserve Banks" in Table 1.{{p}}The Statistical Release H.3, "Aggregate Reserves of Depository Institutions and the Monetary Base," reports reserve balances maintained{{p}}by month and by two-week maintenance period. "Reserve balances maintained/Total" for a given two-week maintenance period on Table{{p}}1 of the H.3 is the average of the two "Reserve balances with Federal Reserve Banks" weekly averages reported in the given{{p}}maintenance period in Table 1 of the H.4.1.{{p}}Interest expense on reserves is calculated based on balances that comprise "Other deposits held by depository institutions" reported in{{p}}Table 5. Information on how interest is calculated on required and excess reserves can be found in the Reserve Maintenance Manual.{{p}}Interest expense on reserves held by depository institutions is presented in the combined statements and quarterly reports. Within these{{p}}reports, interest expense can be found in the Combined Statements of Income and Comprehensive Income (SOI) under the "Interest{{p}} Expense: Deposits: Depository institutions" line of the statements. The expense information is presented inclusive of required and excess{{p}}reserves and does not provide the interest expense related to excess reserves separately.5{{p}}Figure 2{{p}}Figure 3{{p}} FRB: FEDS Notes: Where do I see the Monetary Policy Normalization Tools on the Fed's... Page 2 of 5{{p}} 5/11/2016{{p}}ON RRP offering rate and RRPs{{p}}Another administered rate to help control the level of the effective federal funds rate is the overnight reverse repurchase program (ON{{p}}RRP) offering rate. In general, any counterparty that is eligible to participate in the ON RRP facility should be unwilling to invest funds{{p}}overnight with another counterparty at a rate below the ON RRP rate. Information on the ON RRP operations can be found on the Federal{{p}}Reserve Bank of New York's "Temporary Open Market Operations" website.{{p}}Both Table 1 and Table 5 of the H.4.1 report "Reverse repurchase agreements" under "Factors absorbing reserve balances" and{{p}}"Liabilities," respectively.{{p}}Figure 4{{p}}Figure 5{{p}} FRB: FEDS Notes: Where do I see the Monetary Policy Normalization Tools on the Fed's... Page 3 of 5{{p}} 5/11/2016{{p}}As shown above, Table 1 breaks out RRPs into two categories: "Foreign official and international accounts" and "Primary Dealers and{{p}}expanded counterparties." All RRPs reported in the latter category, include transactions with primary dealers and other financial{{p}}counterparties who have met eligibility criteria to transact in reverse repurchase agreements that serve as an effective tool for managing{{p}}money market interest rates to help support a floor on those rates.6{{p}}In the combined statements and quarterly reports, interest expense information related to RRPs is available in the "System Open Market{{p}} Account: Securities sold under agreements to repurchase" line of the SOI. More detailed information on RRP balances and expense,{{p}}including the break-out of RRP categories equivalent to Table 1 of the H.4.1, is presented in footnote 5 of the combined statements and{{p}}tables 4 and 13 of the quarterly report.{{p}}1. The authors thank Greg Evans, Jane Ihrig, Elizabeth Klee, and Lawrence Mize for comments. Return to text{{p}}2. Table 1 is not the balance sheet, but it is derived primarily from components of the Federal Reserve's balance sheet. Hence, many of the line items reported{{p}}under "Assets" in Table 5 are also reported under "Factors supplying reserve balances," and this parallelism is also evident among line items reported as{{p}}"Liabilities" in Table 5 and "Factors absorbing reserve balances" in Table 1. Return to text{{p}}3. The Fed also raised the primary credit rate when it began raising short-term interest rates. The primary credit rate is the interest rate at which banks can{{p}}borrow reserves overnight from the Federal Reserve. From early 2010 to late 2015, the primary credit rate was set at 75 basis points or 50 basis points above{{p}}the top of the range for the target federal funds rate. Given that reserves are now superabundant and will remain so for some time, depository institutions will not{{p}}need to borrow from the Federal Reserve and so are unlikely to be influenced by the level of the primary credit rate. Return to text{{p}}Figure 6{{p}} FRB: FEDS Notes: Where do I see the Monetary Policy Normalization Tools on the Fed's... Page 4 of 5{{p}} 5/11/2016{{p}}Last update: May 11, 2016{{p}}Home | Economic Research & Data{{p}}4. For a more detailed discussion surrounding the dynamics at play in the federal funds market, please see Ihrig, Meade, and Weinbach. 2015. "Rewriting{{p}}Monetary Policy 101: What's the Fed's Preferred Post-Crisis Approach to Raising Interest Rates?" Journal of Economic Perspective 29 (4): 177-198. Return to{{p}}text{{p}}5. Currently, IOER is set equal to IORR. Therefore, it is possible to approximate interest expense at a higher frequency by multiplying (IOER/365) by the level of{{p}}balances maintained by depository institutions. Return to text{{p}}6. Participation in the ON RRP operations is open to the Federal Reserve's primary dealers as well as its expanded RRP counterparties, which covers a wide{{p}}range of entities including 2a-7 money market funds, banks, and government-sponsored enterprises (Fannie Mae, Freddie Mac, and Federal Home Loan{{p}}Banks). Return to text{{p}}Please cite this note as: Miller, Christian S. and Casey H. Clark (2016). "Where do I see the Monetary Policy Normalization Tools on the{{p}}Fed's Balance Sheet and Income Statement? ," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, May 11,{{p}}2016,{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}} FRB: FEDS Notes: Where do I see the Monetary Policy Normalization Tools on the Fed's... Page 5 of 5{{p}} 5/11/2016
    Date: 2016–05–11
  6. By: Giorgia Piacentino (Olin Business School at Washington Unive); Anjan Thakor (olin school of business); Jason Donaldson (Washington University in St Louis)
    Abstract: This paper develops a theory of banking that is rooted in the evolution of banks from warehouses of commodities and precious goods, which occurred even before the invention of coinage or fiat money. The theory helps to explain why modern banks offer warehousing (custodial and deposit-taking) services within the same institutions that provides lending services and how banks create funding liquidity by creating private money. In our model, the warehouse endogenously becomes a bank because its superior storage technology allows it to enforce loan repayment most effectively. The warehouse makes loans by issuing “fake†warehouse receipts—those not backed by actual deposits— rather than by lending out deposited goods. The model provides a rationale for banks that take deposits, make loans, and have circulating liabilities, even in an environment without risk or asymmetric information. Our analysis provides new perspectives on narrow banking, liquidity ratios and reserves requirements, capital regulation, and monetary policy.
    Date: 2016
  7. By: Benjamin K. Johannsen; Elmar Mertens
    Abstract: Print{{p}}February 9, 2016{{p}}The Expected Real Interest Rate in the Long Run: Time Series Evidence with the Effective{{p}}Lower Bound{{p}}Benjamin K. Johannsen and Elmar Mertens1{{p}}Introduction{{p}}In response to the global financial crisis, the Federal Open Market Committee lowered the target for the federal funds rate to a range of{{p}}0 to 25 basis points in December 2008, and maintained that target range until the end of 2015. Over that same period, longer-term{{p}}interest rates in the United States were at historically low levels. When the Federal Open Market Committee began its policy{{p}}normalization process in December 2015, it stated that "the federal funds rate is likely to remain, for some time, below levels that are{{p}}expected to prevail in the longer run."{{p}}In this note, we present estimates of the expected long-run level of the real federal funds rate, which--together with long-run inflation{{p}}expectations--makes up the level of the nominal federal funds rate that is expected to prevail in the long run. In recent years,{{p}}researchers and commentators have pointed to a possible decline in the longer-run normal level of the real federal funds rate that could{{p}}have been caused by a number of economic factors--such as a decline the trend rate of output growth or an aging population.2 We use a{{p}}statistical framework to estimate the expected long-run normal level of the real federal funds rate. To distinguish cyclical variations in the{{p}}real interest rate from those with longer-run consequences we condition our estimates on a measure of economic slack, inflation, and{{p}}the levels of short- and longer-term interest rates. Since the nominal federal funds rate has been at (or very near) its effective lower{{p}}bound (ELB) in recent years, we embed the concept of so-called shadow rates (further explained below) in a time-series model that{{p}}captures the joint dynamics of interest rates, inflation and economic slack. While we find some supporting evidence for a decline in the{{p}}expected longer-run level of the federal funds rate, our estimates also show that these results are surrounded by large amounts of{{p}}uncertainty.{{p}}Methodology and Data{{p}}We use a dynamic time-series model to characterize the evolution of the following four observable variables: the nominal federal funds{{p}}rate, the nominal yield on Treasury bonds with a maturity of five years, headline PCE inflation, and a measure of the unemployment gap{{p}}based on the Congressional Budget Office's (CBO) estimates of the natural rate of unemployment in the long term.{{p}}A key element of our modeling approach is a decomposition of data into trends and cyclical components. In macroeconomics, there is a{{p}}long history of distinguishing between permanent and transitory effects of economic disturbances (see, for example, Blanchard and{{p}}Fischer, 1989). Specifically, in the spirit of the frequently-used framework of Beveridge and Nelson (1981), we identify the trend{{p}}components in each variable from the long-run forecasts generated by our model for each variable. Assuming that monetary policy can{{p}}affect real variables only temporarily, the trend in real interest rates identified by our model can also be considered a measure of the{{p}}longer-run neutral equilibrium real rate.3{{p}}The following trend components are specified in our model: a common trend in nominal interest rates and an inflation trend; the{{p}}unemployment gap is expected to converge to zero in the long run.4 The common trend assumption for nominal interest rates implies{{p}}that fluctuations in interest rate spreads will eventually peter out without changes in long-term spreads. As a result, our model's long-run{{p}}forecasts for the federal funds rate and the nominal Treasury bond are restricted to move in lockstep.5 Furthermore, appealing to a{{p}}long-run version of the Fisher equation, the trend in the nominal federal funds rate is modeled as the sum of the inflation trend and the{{p}}real-rate trend, which we use to measure the longer-run normal level of the real federal funds rate. As a result, the expected long-run{{p}}level of the nominal federal funds rate can change over time either because of variations in trend inflation or because of changes in the{{p}}longer-term expected real rate.{{p}}The trend component of inflation allows our model to capture the persistent rise of the average level of inflation in the 1970s and its{{p}}subsequent decline. The variability and persistence of inflation has varied considerably over our sample (see for example Cogley,{{p}}Primiceri and Sargent, 2010) and we build on the work by Stock and Watson (2007, 2010, 2015), Garnier, Mertens and Nelson (2015),{{p}}Shephard (2016) and Mertens (forthcoming) by assuming that the volatility of the trend component of inflation--as well as the variability{{p}}of cyclical shocks to inflation as explained in more detail below--varies over time.6{{p}}We model the cyclical components of inflation, the federal funds rate, the long-term interest rate, and the unemployment gap as a{{p}}vector-autoregressive process. In order to capture the notable variations in the size of the business cycle in post-war U.S. data (see{{p}}Bernanke (2004) for a detailed discussion), the gap components of our model are assumed to have time-varying volatility; as in, for{{p}}example, Justiniano and Primiceri (2008). Modeling the cyclical components in a joint vector autoregression captures the co-movements{{p}}and interactions between current and past values of each of the variables in a parsimonious and fairly agnostic way. Among others, this{{p}} FRB: FEDS Notes: The Expected Real Interest Rate in the Long Run: Ti...{{p}}1 of 6 2/10/2016 8:15 AM{{p}}modeling strategy allows the federal funds rate to respond to movements in economic slack--as measured by the unemployment rate{{p}}gap--and cyclical variations of inflation. It also allows the unemployment rate gap to move in conjunction with changes in financial{{p}}conditions that are captured by the cyclical component of the yield on longer-term Treasury bonds or for cyclical fluctuations in inflation{{p}}to depend on changes in the unemployment rate gap.{{p}}Our model is estimated using quarterly data ranging from 1960:Q1 to 2015:Q4; all data is publicly available from the FRED database{{p}}maintained by the Federal Reserve Bank of St. Louis.7 Inflation is measured by the quarterly rate of change in the PCE headline deflator{{p}}(expressed as an annualized percentage rate).8 Readings for the federal funds rate and the 5-year nominal bond yields are constructed{{p}}as quarterly averages of the effective federal funds rate and the Treasury's 5-year constant maturity rate, respectively.9 The{{p}}unemployment gap is computed as the difference between the quarterly average rate of unemployment and the CBO's measure of the{{p}}natural long-term rate of unemployment for a given quarter.10 All computations are based on the vintage of FRED data available that has{{p}}been available at the end of January 2016.{{p}}Modeling Interest Rates at the Effective Lower Bound{{p}}A challenge for our statistical model is that, from late 2008 to late 2015, the federal funds rate was at (or very near) the ELB. This lower{{p}}bound on nominal interest rates potentially alters the co-movement between the federal funds rate and the other data in our sample.{{p}}Additionally, the non-linearity of the ELB is often ignored by traditional time-series methods. To overcome this technical challenge, we{{p}}define the "shadow" federal funds rate, which is a notional rate that is constructed to be less than the zero during the period that the{{p}}target range for the federal funds rate was between 0 and 25 basis points. During all other times, the shadow rate is assumed to equal to{{p}}the federal funds rate. In this way, the observed nominal federal funds rate can be thought of as a censored version of the shadow rate.{{p}}At the ELB we do not observe the shadow rate, but estimates of the shadow rate can be generated based on the historical persistence{{p}}and co-movements between the federal funds rate and the other series in our model.{{p}}Several other studies have used the term "shadow rate" to refer to a similar modeling object, notably Krippner (2013), Wu and Xia{{p}}(2015), and Bauer and Rudebusch (2014) that is typically derived from no-arbitrage conditions embedded in a term structure model for{{p}}nominal interest rates. Our contribution is that we estimate time-variation in the longer-run normal level of the federal funds rate in a{{p}}somewhat more agnostic time-series framework that infers the shadow rate merely by treating nominal interest rates at the ELB as{{p}}censored shadow rates, but without imposing any particular economic structure. As a first step towards our approach the federal funds{{p}}rate at the ELB could also be treated as missing data (without regard for any censoring constraints), which is equivalent to allowing the{{p}}shadow rate to take any value during that period, as has been done, for example, by Tallman and Zaman (2012). In this case, the{{p}}information contained in the federal funds rate during the period of the ELB would be discarded and the shadow rate would be inferred{{p}}merely from the historical co-movements between the federal funds rate and the other macroeconomic variables used in our model.{{p}}Below, we present results from both modeling approaches.{{p}}Estimates of the Longer-Run Level of the Real Federal Funds Rate{{p}}Figure 1 shows our estimate of the expected long-run real federal funds rate when we impose the restriction that the shadow rate be{{p}}less than the zero during the period of the ELB. Panel A shows estimates that, at each date, use information up to that date (we refer to{{p}}these as filtered estimates). Panel B shows estimates that use the entire data sample (we refer to these as smoothed estimates).{{p}}Nevertheless, as the figure shows, we find some evidence of a decline in the expected longer-run federal funds rate using either the{{p}}filtered or the smoothed estimates.{{p}}Figure 1: Expected Long-Run Real Federal Funds Rate{{p}} Note: At each date, filtered estimates use data up through that date to estimate the parameters of the model and the level of the long-run real federal funds{{p}}rate. Smoothed estimates use data through the entire sample to estimate the parameters and the historical level of the long-run real federal funds rate. For the{{p}}results in this figure, we impose that the shadow rate be less than zero during the period of the ELB. Shaded regions are 50 and 90 percent uncertainty bands.{{p}}Estimated trends are expressed in annual percentage terms.{{p}} FRB: FEDS Notes: The Expected Real Interest Rate in the Long Run: Ti...{{p}}2 of 6 2/10/2016 8:15 AM{{p}}Accessible Version{{p}}The uncertainty bands surrounding our estimates are wide, indicating uncertainty about both the level of and changes in the expected{{p}}longer-run real federal funds rate is large. These results are reminiscent of results reported by Hamilton and others (2015), Kiley (2015),{{p}}and Lubik and Matthes (2015) who also report large uncertainty bands surrounding estimates of the longer-run real rate.{{p}}An important feature of our estimated trend in real interest rates is that any decline began well-before the onset of the Great Recession;{{p}}in particular also when considering our filtered estimates that do not condition on future observations from our data set. Our model sees{{p}}the low federal funds rate since 2008 through the lens of co-movement among all of the variables in our model. Moreover, the{{p}}time-varying volatility of the gap components helps the model attribute large changes in unemployment and inflation, like those seen at{{p}}the onset of the global financial crisis, to cyclical fluctuations. This flexibility of our model helps explain why our model sees less{{p}}movement in the trend real interest rate over the past decade than results reported in some other studies, for example Laubach and{{p}}Williams (2015). Instead, the model delivers an estimate of a slow-moving and long-lived decline in the real rate trend that appears to{{p}}have continued through the past few years.{{p}}To ensure that our results are not overly influenced by our shadow rate modeling device, we also conduct our analysis assuming that the{{p}}federal funds rate data are missing whenever the federal funds rate average for a quarter was less than 25 basis points. Figure 2 shows{{p}}the estimated real rate trend under this alternative treatment of the data. As in Figure 1, panels A and B show filtered and smoothed{{p}}estimates. We again find that the estimated real rate trend is surrounded by lots of uncertainty and that any estimated decline in the{{p}}trend began well before the onset of the global financial crisis.{{p}}Figure 2: Expected Long-Run Real Federal Funds Rate, No ELB{{p}} Note: At each date, filtered estimates use data up through that date to estimate the parameters of the model and the level of the long-run real federal funds{{p}}rate. Smoothed estimates use data through the entire sample to estimate the parameters and the historical level of the long-run real federal funds rate. For the{{p}}results in this figure, we do not impose that the shadow rate be less than zero during the period of the ELB. Shaded regions are 50 and 90 percent uncertainty{{p}}bands. Estimated trends are expressed in annual percentage terms.{{p}}Accessible Version{{p}}Figure 3 shows our estimates of the shadow rate. Panel A shows the shadow rate when we impose the censoring constraint which{{p}}requires that the shadow rate be less than zero during the ELB period and panel B shows the shadow rate when we do not impose this{{p}}restriction. While the estimated shadow rates from the two procedures are not identical, they show similar movements over time. This{{p}}indicates that the information contained in the longer-term interest rate, as well as the information from inflation and the unemployment{{p}}rate gap, is informative about shorter-term rates, and that our shadow rate modeling device is unlikely to be driving our results regarding{{p}}the longer-run real federal funds rate.{{p}}Figure 3: Estimated Shadow Rate{{p}} FRB: FEDS Notes: The Expected Real Interest Rate in the Long Run: Ti...{{p}}3 of 6 2/10/2016 8:15 AM{{p}} Note: At each date, the estimates of the shadow rate shown in both panels use data through the entire sample to estimate the parameters and the historical{{p}}level of the shadow rate. Prior to 2008, the shadow rate in our model is equal to the federal funds rate. Shaded regions are 50 and 90 percent uncertainty{{p}}bands. Estimated shadow rates are expressed in annual percentage terms.{{p}}Accessible Version{{p}}Our final figure displays the data we use as inputs to our model estimation, along with the smoothed trends that our model produces.{{p}}Figure 4: Macroeconomic Data and Estimated Trends{{p}} Note: Raw data are shown as the red solid line. At each date, we use data through the entire sample to estimate the parameters and the historical levels of the{{p}}trends shown. Shaded regions are 50 and 90 percent uncertainty bands. Inflation and interest rates, along with their estimated trends, are expressed in annual{{p}}percentage terms. Vertical dotted lines indicate peaks and troughs associated with recession, as dated by the National Bureau of Economic Research,{{p}} .{{p}}Accessible Version{{p}}Conclusions{{p}}Interest rates in the United State have been historically low since 2008. Given the severity of the recession that followed the global{{p}}financial crisis, it is no surprise that interest rates fell. However, it remains to be seen how much of that decline will be long-lasting.{{p}}Our modeling framework allows us to estimate the trend component of the real federal funds rate while filtering out cyclical fluctuations in{{p}} FRB: FEDS Notes: The Expected Real Interest Rate in the Long Run: Ti...{{p}}4 of 6 2/10/2016 8:15 AM{{p}}the real federal funds rate based on its co-movement with other macroeconomic conditions as measured by the CBO's unemployment{{p}}rate gap, inflation, and financial conditions represented by the level of longer-term Treasury yields. As our exercise draws on nominal{{p}}interest rate data, we also account for the period of the ELB by modeling observed nominal interest rates as censored realizations of{{p}}so-called shadow rates, which are hypothetical nominal interest rates implied by our model for the case where interest rates could fall{{p}}below ELB. Our results suggest that any decline in the trend component of real interest rates over recent years can be thought of as a{{p}}continuation of a decline that began years before. However, our results also suggest that estimates of the longer-run expected value of{{p}}the federal funds rate are surrounded by a large amount of uncertainty.{{p}}References{{p}}Bauer, Michael and Glenn Rudebusch, forthcoming, "Monetary Policy Expectations at the Zero Lower Bound," Journal of Money, Credit,{{p}}and Banking.{{p}}Bernanke, Ben S., February 20, 2004, "The Great Moderation," Remarks at the meetings of the Eastern Economic Association,{{p}}Washington, DC.{{p}}Beveridge, Stephen and Charles R. Nelson, 1981, "A New Approach to Decomposition of Economic Time Series into Permanent and{{p}}Transitory Components with Particular Attention to Measurement of the 'Business Cycle'," Journal of Monetary Economics, 7(2), pp.{{p}}151-174.{{p}}Blanchard, Olivier and Stanley Fischer, 1989, March, "Lectures on Macroeconomics," MIT Press, ISBN: 9780262022835.{{p}}Cogley, Timothy, Giorgio E. Primiceri, and Thomas J. Sargent, 2010. "Inflation-Gap Persistence in the US," American Economic Journal:{{p}}Macroeconomics, 2(1): 43-69.{{p}}Garnier, Christine, Elmar Mertens, and Edward Nelson, 2015, September, "Trend Inflation in Advanced Economies," International{{p}}Journal of Central Banking.{{p}}Hamilton, James D., Ethan S. Harris, Jan Hatzius, and Kenneth D. West, 2015, October, "The Equilibrium Real Funds Rate: Past,{{p}}Present, and Future," Hutchins Center on Fiscal & Monetary Policy at Brookings Working Paper #16.{{p}}Kiley, Michael T., 2015, August, "What Can the Data Tell Us About the Equilibrium Real Interest Rate?" Finance and Economics{{p}}Discussion Series, Federal Reserve Board of Governors, 2015-077.{{p}}Krippner, Leo, 2013, August, "A Tractable Framework for Zero Lower Bound Gaussian Term Structure Models," Reserve Bank of New{{p}}Zealand Discussion Paper Series, DP2013/02.{{p}}Laubach, Thomas and John C. Williams, 2015, November, "Measuring the Natural Rate of Interest Redux," Hutchins Center on Fiscal &{{p}}Monetary Policy at Brookings Working Paper #15.{{p}}Lubik, Thomas A. and Christian Matthes, 2015, October. "Calculating the Natural Rate of Interest: A Comparison of Two Alternative{{p}}Approaches." Economic Brief No. 15-10. Federal Reserve Bank of Richmond.{{p}}Mertens, Elmar, forthcoming, "Measuring the Level and Uncertainty of Trend Inflation," Review of Economics and Statistics.{{p}}Rachel, Lukasz and Thomas D. Smith, 2015, December, "Secular Drivers of the Global Real Interest Rate," Bank of England Staff{{p}}Working Paper, No. 571.{{p}}Shephard, Neil, 2016. "Martingale Unobserved Component Models," in Unobserved Components and Time Series Econometrics edited{{p}}by Siem Jan Koopman and Neil Shephard, Oxford University Press.{{p}}Stock, James H. and Mark W. Watson, 2007, February, "Why has U.S. Inflation Become Harder to Forecast?" Journal of Money, Credit,{{p}}and Banking 29(S1), 3-33.{{p}}Stock, James H. and Mark W. Watson, 2010, October, "Modeling Inflation after the Crisis" NBER Working Papers 16488, National{{p}}Bureau of Economic Research.{{p}}Stock, James H. and Mark W. Watson, 2015, June, "Core Inflation and Trend Inflation," NBER Working Papers 21282, National Bureau{{p}}of Economic Research.{{p}}Summers, Lawrence H., 2014, "US Economic Prospects: Secular Stagnation, Hysteresis and the Zero Lower Bound," Business{{p}}Economics, 49(2), pp. 65-73.{{p}}Tallman, Ellis W. and Saeed Zaman, 2012, October. "Where Would the Federal Funds Rate Be, If It Could Be Negative?" Economic{{p}}Commentary No. 2012-15, Federal Reserve Bank of Cleveland.{{p}}Wu, Jing Cynthia and Fan Dora Xia, forthcoming, "Measuring the Macroeconomic Impact of Monetary Policy at the Zero Lower Bound,"{{p}}Journal of Money, Credit, and Banking.{{p}}Yellen, Janet L., December 2, 2015. "The Economic Outlook and Monetary Policy," Remarks at the Economic Club of Washington, D.C.{{p}}1. Board of Governors of the Federal Reserve System. The views expressed here are those of the authors and not necessarily the views of the Board of{{p}}Governors, the FOMC, or anyone else associated with the Federal Reserve System. Return to text{{p}} FRB: FEDS Notes: The Expected Real Interest Rate in the Long Run: Ti...{{p}}5 of 6 2/10/2016 8:15 AM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}2. See, for example, Summers (2014) and Rachel and Smith (2015). Return to text{{p}}3. The neutral equilibrium real rate is typically defined as the inflation-adjusted "value of the federal funds rate that would be neither expansionary nor{{p}}contractionary if the economy were operating near its potential" (Yellen, 2015). Since our estimated real-rate trend can only provide a perspective on{{p}}longer-run expectations of the neutral equilibrium real rate, it does not provide an indication about the appropriate stance of monetary policy in the near term.{{p}}Return to text{{p}}4. Implicitly, our model treats the CBO's estimate of the natural rate of unemployment as the (known) trend rate of unemployment. Return to text{{p}}5. The common trend assumption does not restrict long-run forecasts of short- and long-term interest rates to be identical because we additionally estimate an{{p}}average difference between long-term and short-term rates. Return to text{{p}}6. Specifically, we allow for time-varying volatility in three components of the inflation process: trend and cycle plus a serially uncorrelated measurement error{{p}}that serves to filter out the high-frequency variations in headline inflation. Return to text{{p}}7. See . Return to text{{p}}8. See . Return to text{{p}}9. Available at , and , respectively. Return to text{{p}}10. Available at and . Return to text{{p}}Please cite this note as:{{p}}Benjamin K. Johannsen and Elmar Mertens (2016). "The Expected Real Interest Rate in the Long Run: Time Series Evidence with the{{p}}Effective Lower Bound," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, February 9,{{p}}/10.17016/2380-7172.1703.{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: February 9, 2016{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: The Expected Real Interest Rate in the Long Run: Ti...{{p}}6 of 6 2/10/2016 8:15 AM
    Date: 2016–02–09
  8. By: Alain Kabundi; NtuthukoTsokodibane
    Abstract: With the adoption of the in‡ation targeting (IT) regime in 2000, the South African Reserve Bank (SARB) became independent. With the independence of monetary policy comes accountability to the public at large, which in turn leads to transparency in the conduct of monetary policy. The SARB has come a long way in its communication strategy. In 2014 it adds another layer in its communication strategy by announcing explicitly throughout 2014 that monetary policy was on the rising cycle until normalisation is reached. Monetary policy committee (MPC) statements of March and May 2014 refer to normalisation as the return of the policy rate (repo rate) from the historical lowest level of 5 per cent to the normal level in the long run. Like many central banks, the SARB reduced the policy rate from 12 per cent to 5 per cent following the Global Financial Crisis (GFC).
    Keywords: are Monetary Policy, Central Bank Communication, and Nonparametric Change Point.
    JEL: C14 E43 E52 E58 G14
    Date: 2016–08
  9. By: Seneca, Martin (Bank of England)
    Abstract: Large risk shocks give rise to cost-push effects in the canonical New Keynesian model. At the same time, monetary policy becomes less effective. Therefore, stochastic volatility introduces occasional trade-offs for monetary policy between inflation and output gap stabilisation. The cost-push effects operate through expectational responses to the interaction between shock volatility and the zero lower bound (ZLB) on interest rates. Optimal monetary policy calls for potentially sharp reductions in the interest rate when risk is elevated, even if this risk never materialises. Close to the ZLB, small risk shocks become ‘large’ in this sense. If policy is initially constrained by the ZLB, lift-off is optimally delayed when risk increases.
    Keywords: Risk shocks; uncertainty; zero lower bound on interest rates; optimal monetary policy
    JEL: E52 E58
    Date: 2016–08–11
  10. By: Ellen E. Meade; Miguel Acosta
    Abstract: Print{{p}}Chart 1. FOMC statements, word count by meeting{{p}}September 30, 2015{{p}}Hanging on every word: Semantic analysis of the FOMC's postmeeting statement{{p}}Ellen Meade, Miguel Acosta1{{p}}Introduction{{p}}The Federal Open Market Committee's (FOMC or "Committee") postmeeting statements constitute one of the key vehicles through{{p}}which the Committee communicates its assessment of the economy, its policy actions, and its thinking about future policy. In this note,{{p}}we use techniques developed in natural language processing (NLP, also referred to as computational linguistics) to study how the{{p}}content of the FOMC's postmeeting statement has changed from May 1999--when the Committee began releasing statements{{p}}consistently--through December 2014. We show that, to the untutored reader, the statements appear to be highly similar from one{{p}}FOMC meeting to the next. Indeed, the semantic persistence of the statements has risen notably in recent years relative to the early{{p}}years of our sample period. Importantly, we find that, after using some tools commonly employed in computational linguistics to prepare{{p}}the text, the semantic persistence of FOMC statements from one meeting to the next is revealed to be much lower than casual reading{{p}}would suggest and is also more variable. We see this as evidence of a greater information content present in the postmeeting statement{{p}}than might be apparent at a first glance. We highlight the December 2008 statement--an historically significant statement because of the{{p}}policy actions taken at that FOMC meeting--to illustrate the more general point that NLP techniques offer a new lens through which to{{p}}view the Federal Reserve's communications. Our analysis shows that natural language processing can strip away false impressions and{{p}}uncover hidden truths about complex communications such as those of the Federal Reserve.{{p}}Background on the FOMC statement{{p}}More than two decades ago, in February 1994, Chairman Greenspan issued the first postmeeting statement following the FOMC's{{p}}decision to tighten monetary policy--the first increase in the target for the federal funds rate since 1989. For the next five years, a{{p}}statement was released only after FOMC meetings at which the Committee changed its target for the federal funds rate. In May 1999,{{p}}the FOMC began releasing statements at the conclusion of every meeting regardless of whether it had decided to make a change to the{{p}}policy interest rate or not. Later, in May 2002, the FOMC decided to provide voting information in the postmeeting statements, including{{p}}a short explanation of any dissenting votes, thereby accelerating the provision of this information to the public by several weeks.2{{p}}The FOMC's statements have become substantially longer and more informative over time. The first statement issued in February 1994{{p}}was a mere 99 words in 4 sentences, compared with 564 words in 22 sentences in December 2014.3 Chart 1 provides the word count{{p}}for FOMC statements since May 1999. Hernández-Murillo and Shell (2014)4 analyze the linguistic complexity of the FOMC statement{{p}}using the well-known Flesch-Kincaid Grade Level index, which measures the reading grade level of a particular text. They find that,{{p}}while the early statements are written at a reading grade level of 9 to 14 years of schooling, the more recent statements are written at a{{p}}reading grade level of three years beyond a 4-year college degree.{{p}}Early{{p}}statements{{p}}included a{{p}}brief{{p}}economic{{p}}rationale for{{p}}the change{{p}}in monetary{{p}}policy; more{{p}}recent{{p}}statements{{p}}include an{{p}}assessment{{p}}of current{{p}}economic{{p}}conditions{{p}}and a{{p}}discussion{{p}}of the{{p}}economic{{p}}outlook, as{{p}}well as the{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}1 of 9 9/30/2015 3:56 PM{{p}} Note: Excludes voting information{{p}}Accessible version{{p}}Chart 2. Selected word counts from FOMC statements{{p}}rationale for{{p}}current{{p}}monetary{{p}}policy and{{p}}forward{{p}}guidance{{p}}about the{{p}}evolution of{{p}}policy in the{{p}}future. Chart{{p}}2 provides{{p}}some{{p}}word-count-based evidence about the references to specific topics in the statements over time. For example, the weather and global{{p}}developments have been mentioned in the statement relatively infrequently; movements in energy or other commodity prices have{{p}}received greater attention, as has growth in productivity, which figured frequently in statements prior to mid-2006. In contrast, references{{p}}to inflation expectations have become more frequent since the financial crisis with at least one reference in each statement since{{p}}September 2009.5{{p}}Measuring{{p}}semantic{{p}}similarity{{p}}Recently,{{p}}economists{{p}}have begun{{p}}to use{{p}}techniques{{p}}from the{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}2 of 9 9/30/2015 3:56 PM{{p}}Accessible version{{p}}computational linguistics literature that have been employed extensively in political science and other disciplines to examine complex{{p}}texts.6 Fed watchers and others in the public who pay close attention to the Committee's statements commonly examine a "tracked{{p}}changes" version of the statement upon its release in order to look closely at words that have been added or dropped for clues about{{p}}changes in the Committee's views on the economy or its policy intentions. Here, the tools we employ measure one aspect that is closely{{p}}related to that "tracked changes" approach: the correlation of words used in two consecutive postmeeting statements.{{p}}More formally, we compute the "cosine similarity" between the "vector-space models" of consecutive FOMC statements to tell us how{{p}}persistent the content of the statements has been over time. First, we create a 1302-by-126 matrix in which each row corresponds to{{p}}one of the 1302 unique words contained in our sample of FOMC statements and each column corresponds to a single statement. We{{p}}then populate the cells of the matrix with the frequency of each word in each statement making each column of the matrix a{{p}}vector-space model of the corresponding FOMC statement. The cosine similarity between two FOMC statements, a and b, is equal to:{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}3 of 9 9/30/2015 3:56 PM{{p}}Table 1. Tracking changes in the December 2008 FOMC statement{{p}}(Strikeout shows language that appeared in the October 2008 postmeeting statement but did not appear in the December{{p}}2008 statement; red language appeared in the December statement but not in the October statement; and words in black{{p}}but not strikeout appeared in both statements.){{p}}where n is the number of unique words (1302 in this case); ai and bi represent the number of times that word i occurs in statements a{{p}}and b, respectively. From this definition, we see that two unrelated or "orthogonal" documents will have a cosine similarity equal to zero{{p}}because they share no words (one or both of ai or bi equals zero, for all i).7 Documents that use the same words in nearly the same{{p}}proportion will have a cosine similarity that is close to unity. Thus, cosine similarity in NLP measures something very close to a{{p}}correlation coefficient in data analysis. Measuring the cosine similarity for each pair of consecutive statements permits us to examine{{p}}how consistent word usage is from statement to statement, with values close to unity representing high semantic similarity and{{p}}persistence.{{p}}Semantic similarity of raw postmeeting statements{{p}}In this section, we examine the cosine similarity of the raw postmeeting statements. By "raw," we mean that the statements have not{{p}}been subjected to any of the preprocessing that is common in natural language processing. That is, after downloading the statements{{p}}from the Federal Reserve Board of Governors web site, we removed only punctuation, the paragraph(s) that provides voting information{{p}}and an explanation of dissenting votes, if any, and the paragraphs at the end of some statements that report changes to the discount{{p}}rate approved by the Board of Governors.{{p}}Returning to the "tracked changes" analogy, if the same words are used in consecutive FOMC statements (and we ignore changes in{{p}}word order), we will see no red ink--that is, no changes--and cosine similarity will equal unity. The addition or subtraction of words or the{{p}}use of the same words in different proportions will cause us to see some red ink and thus reduce cosine similarity. We illustrate this in{{p}}table 1 with a tracked-changes version of the FOMC statement from December 2008. At that meeting, the Committee reduced its target{{p}}for the federal funds rate to a range of 0 to 1/4 percent, noted in paragraph 1 of the statement, and added a lengthy new fifth paragraph{{p}}discussing the Federal Reserve's balance sheet and asset purchase programs. The Committee also updated its views of economic{{p}}activity and inflation in paragraphs 2 and 3. Significant policy changes were undertaken at the December 2008 meeting; such large{{p}}changes in the statement language between the October and December 2008 meetings would be expected to result in a relatively low{{p}}value of cosine similarity.{{p}}Chart 3{{p}}shows the{{p}}cosine{{p}}similarity of{{p}}consecutive,{{p}}raw FOMC{{p}}statements{{p}}(the blue{{p}}line) and an{{p}}8-meeting{{p}}moving{{p}}average (the{{p}}red line);8{{p}}note that{{p}}cosine{{p}}similarity{{p}}dropped{{p}}sharply in{{p}}December{{p}}2008 to{{p}}about 0.8,{{p}}as predicted{{p}}by our{{p}}analysis of{{p}}the tracked{{p}}changes in{{p}}the{{p}}statement.{{p}}Over the{{p}}entire{{p}}sample{{p}}period, the{{p}}meeting-to-{{p}}meeting{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}4 of 9 9/30/2015 3:56 PM{{p}}Accessible version{{p}}Chart 3. Semantic similarity of consecutive, raw FOMC statements{{p}}persistence{{p}}in the raw{{p}}statements{{p}}has{{p}}averaged{{p}}0.93.9{{p}}Between the{{p}}start of the{{p}}sample in{{p}}May 1999{{p}}and{{p}}mid-2003,{{p}}persistence{{p}}averaged{{p}}0.86, before{{p}}increasing to{{p}}0.94{{p}}between{{p}}mid-2003{{p}}and{{p}}mid-2007.{{p}}Average{{p}}persistence{{p}}declined{{p}}during the{{p}}financial{{p}}crisis and{{p}}then rose to{{p}}a very high{{p}}level{{p}}between{{p}}2009 and{{p}}2014. In{{p}}addition,{{p}}linguistic{{p}}persistence{{p}}was quite{{p}}variable until{{p}}about{{p}}mid-2009, but has shown little meeting-to-meeting variation since then.{{p}}Preparation{{p}}of the text{{p}}using{{p}}standard{{p}}tools from{{p}}natural{{p}}language{{p}}processing{{p}}While cosine{{p}}similarity of{{p}}the raw{{p}}documents{{p}}provides a{{p}}simple{{p}}illustration of{{p}}the semantic{{p}}correlation{{p}}of FOMC{{p}}statements{{p}}from{{p}}meeting to{{p}}meeting, it{{p}}may not{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}5 of 9 9/30/2015 3:56 PM{{p}}Accessible version{{p}}Table 2. Effects of text preparation{{p}}Original word or{{p}}term{{p}}After stemming or{{p}}concatenation{{p}}Words that share{{p}}stem{{p}}Total occurrences of{{p}}stem{{p}}Number of documents{{p}}containing stem{{p}}TFIDF{{p}}score{{p}}decisionmaking decisionmak -- 1 1 4.83{{p}}inflation infl inflationary 548 114 0.1{{p}}federal funds rate fedfundsrate -- 188 125 0.08{{p}}Chart 4. Computing the TFIDF score{{p}}accurately{{p}}represent{{p}}the similarity{{p}}of the{{p}}intended or{{p}}understood{{p}}semantic{{p}}content,{{p}}which also{{p}}depends on{{p}}word{{p}}complexity,{{p}}multiple{{p}}meanings of{{p}}the same{{p}}words, and{{p}}different{{p}}variations of{{p}}the same{{p}}root words.{{p}}To get a{{p}}handle on{{p}}these{{p}}issues, we{{p}}employ{{p}}some simple{{p}}tools that{{p}}are commonly used in computational linguistics to prepare-- or "preprocess"--the text; computing cosine similarities after such{{p}}preparation provides a more meaningful and useful interpretation of how persistent the FOMC statements have been over time.{{p}}We subject each statement to three standard preprocessing steps: First, we remove common words that provide little semantic content,{{p}}including pronouns, articles, conjunctions, dates, numbers, and geographic words.10 Next, we concatenate phrases when words used{{p}}together take on a particular meaning--for example, Federal Open Market Committee or federal funds rate--so that different uses of the{{p}}component words can be distinguished. In the final preprocessing step, we "stem" all words down to a root, meaning, for example, that{{p}}increase, increased, increases and increasing are all shortened to increas--the original words are removed from the term-document{{p}}matrix and a row for each stem is added that contains the sum of the counts of the original words. Table 2 provides some examples of{{p}}the effects of the preprocessing steps for particular words. After the preprocessing is completed, the term-document matrix contains 721{{p}}rows (one for each stemmed term) and 126 columns (one for each FOMC statement).{{p}}Finally, we apply a standard weighting scheme known as term frequency–inverse document frequency (TFIDF) to the now-smaller{{p}}term-document matrix. Again, in the term-document matrix, the ijth entry represents the number of times that the stemmed term i is used{{p}}in document j. We compute the number of documents in which term i occurs, ni , and then weight each row--that is, each word--of the{{p}}matrix by ln(n/ni). The effect of this procedure is to give a lower weight to terms that occur in many documents.{{p}}The final column of table 2 lists the TFIDF score for each stemmed term. To better illustrate how the weighting works, chart 4 provides a{{p}}histogram of the number of documents in which each term is used, ni (the blue bars), and the TFIDF score (the black line) associated{{p}}with each value of ni. The blue bars at the right hand side of the graph represent words that occur often in FOMC statements, and thus{{p}}receive a lower weight. For instance, words that occur in every postmeeting statement, such as "Committee," have a TFIDF score of{{p}}zero because such words don't help to distinguish semantic content between documents. Of course, there are meaningful terms--such{{p}}as "inflation" or "federal funds rate"--that also have a very low TFIDF score. In such cases, the context in which the term is used is what{{p}}allows us to assign meaning. And, the rarer are these context words, the higher is their TFIDF weight.{{p}}Semantic{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}6 of 9 9/30/2015 3:56 PM{{p}}Accessible version{{p}}Chart 5.a. Semantic similarity of consecutive, prepared FOMC statements{{p}} Note: Text preprocessing only.{{p}}similarity of{{p}}prepared{{p}}FOMC{{p}}statements{{p}}Charts 5.a{{p}}and 5.b{{p}}illustrate the{{p}}effects of{{p}}the text{{p}}preparation{{p}}steps on the{{p}}cosine{{p}}similarity of{{p}}consecutive{{p}}FOMC{{p}} statements:{{p}}5.a folds in{{p}}the three{{p}}text{{p}}preprocessing steps while 5.b includes both the preprocessing steps and the weighting scheme. The preparation of the text reduces the{{p}}level of cosine similarity, though the upward trend and overall shape of persistence is similar to our baseline case. However, for the{{p}}treated documents there is a level shift downward from the baseline case, which tells us that the postmeeting statements are less similar{{p}}when closer attention is paid to their semantic content. This is especially so when the TFIDF weighting is included, as this weighting is{{p}}based on the importance of individual words over the entire sample of FOMC statements. This is most readily seen in chart 6, which{{p}}shows the moving average of cosine similarity from the baseline case with no preprocessing (the solid line), the moving average when{{p}}text preprocessing is included (the dash-dotted line), and the moving average when both the preprocessing steps and TFIDF weighting{{p}}are included (the dashed line). Here it is clear that cosine similarity with text preparation is more volatile than the baseline and that there{{p}}is lower semantic similarity than in the baseline, particularly when the TFIDF weighting is included.{{p}}Earlier we{{p}}noted that{{p}}the{{p}}December{{p}}2008{{p}}statement{{p}}had a low{{p}}value of{{p}}cosine{{p}}similarity{{p}}(0.8) when{{p}}compared{{p}}with the{{p}}statement{{p}}issued in{{p}}October{{p}}2008. With{{p}}text{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}7 of 9 9/30/2015 3:56 PM{{p}}Accessible version{{p}}Chart 5.b. Semantic similarity of consecutive, prepared FOMC statements{{p}} Note: Text preprocessing and TFIDF weighting.{{p}}Accessible version{{p}}Chart 6. Semantic similarity moving averages{{p}}Accessible version{{p}}preprocessing, cosine similarity drops substantially--to about 0.5--largely because of the removal of commonly-used words and the fact{{p}}that the December statement contained many words not present in the October statement (increasing the denominator relative to the{{p}}numerator in the computation of cosine similarity). Furthermore, when the TFIDF weighting is included, the cosine similarity value of the{{p}}December statement falls to almost 0.1 giving it the lowest score in terms of meeting-to-meeting semantic persistence of any statement{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}8 of 9 9/30/2015 3:56 PM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}in chart 5.b. This tells us that the words introduced in the December statement were also relatively rare--thus, the TFIDF weighting gives{{p}}more prominence to them (increasing the denominator in the cosine similarity computation even further). For example, the terms "sheet"{{p}}(as in "balance sheet"), "benefit," and "est" (the root of "establish") are highly unusual and present in the December statement (but{{p}}absent from the October one).{{p}}Conclusion{{p}}The FOMC postmeeting statement is one of the Federal Reserve's most important communications vehicles. In this note, we have used{{p}}techniques developed in computational linguistics to demonstrate that FOMC statements have become steadily more similar in content{{p}}from meeting to meeting, particularly since the financial crisis. Moreover, once we look beyond the raw language in the statements, we{{p}}find that the semantic content of the statements from one FOMC meeting to the next is less similar and more variable than would appear{{p}}at first glance. We focused on the postmeeting statement that the FOMC issued in December 2008, which announced significant{{p}}changes in monetary policy, illustrating that NLP tools can help us to uncover just how unusual the FOMC statement issued after that{{p}}meeting was and suggesting how useful these tools can be in studying FOMC communications more generally.{{p}}1. The authors thank Steve Meyer and Robert Tetlow for helpful comments. Return to text{{p}}2. The voting information had previously been included only in the meeting minutes, which at that time were released about six weeks after an FOMC meeting.{{p}}Return to text{{p}}3. In order to make the December 2014 statement comparable to the February 1994 one, we stripped out the voting information from the former before taking{{p}}the word and sentence counts. Return to text{{p}}4. Hernández-Murillo, Rubén and Hannah G. Shell (2014), "Rising Complexity of the FOMC Statement," Economic Synopses, Number 23. Return to text{{p}}5. Each panel in Chart 2 shows how many times the given words ("productive" and "productivity," for example) are used in each postmeeting statement.{{p}}Because there are eight FOMC meetings per year, there are eight data points for each set of words each year. Some panels also show a related{{p}}macroeconomic data series from FRED. Return to text{{p}}6. For a discussion of techniques, see David Bholat, Stephen Hansen, Pedro Santos and Cheryl Schonhardt-Bailey (2015), "Text mining for central banks{{p}}(PDF)," CCBS Handbook No. 33, Bank of England; for application of the techniques to central bank texts, see Miguel Acosta (2015), "FOMC Responses to{{p}}Calls for Transparency," Finance and Economic Discussion Series 2015-060, Board of Governors of the Federal Reserve System; Stephen Hansen, Michael{{p}}McMahon, and Andrea Prat (2014), "Transparency and Deliberation within the FOMC: a Computational Linguistics Approach," CEP Discussion Papers DP{{p}}1276, Centre for Economic Performance, London School of Economics; Cheryl Schonhardt-Bailey (2013), Deliberating American Monetary Policy: A Textual{{p}}Analysis, MIT Press, Cambridge, MA. Return to text{{p}}7. Note that the cosine similarity measure does not increase or reduce the similarity of two documents based on differences in document length--this is{{p}}because of the normalization that occurs in the denominator. See the discussion in Bholat et al. (2015). Return to text{{p}}8. There are eight FOMC meetings per year, so an 8-meeting moving average is essentially a 1-year window. Return to text{{p}}9. Taking the average of cosine similarity through 2008 allows for a comparison with measures of FOMC minutes and transcript persistence presented in{{p}}Acosta (2015): FOMC statement persistence from 1999 through 2008 (0.90) is just slightly higher than the raw persistence of FOMC minutes (0.85) or{{p}}transcripts (0.89). Return to text{{p}}10. For example, the word "the" occurs 2,629 times in the statements in our sample. Common text preparation includes the removal of all numbers, a{{p}}procedure we follow here; however, certain numbers--such as the Committee's target for the federal funds rate or the dates in its calendar-based forward{{p}}guidance, may be important to retain. See Bill McDonald's Textual Analysis webpage for the lists used here: Generic; Dates and Numbers; and Geographic. In{{p}}natural language processing, parts of speech such as pronouns, articles, and conjunctions are termed "stop words." Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: September 30, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: Hanging on every word: Semantic analysis of the ...{{p}}9 of 9 9/30/2015 3:56 PM
    Date: 2015–09–30
  11. By: Aikman, David (Bank of England); Bush, Oliver (London School of Economics); Davis, Alan (University of California)
    Abstract: We have entered a world of conjoined monetary and macroprudential policies. But can they function smoothly in tandem, and with what effects? Since this policy cocktail has not been seen for decades, the empirical evidence is almost non-existent. We can only fix this shortcoming in a historical laboratory. The Radcliffe Report (1959), notoriously sceptical about the efficacy of monetary policy, embodied views which led the United Kingdom to a three-decade experiment of using credit controls alongside conventional changes in the central bank interest rate. These non-price tools are similar to policies now being considered or used by macroprudential policymakers. We describe these tools, document how they were used by the authorities, and craft a new, largely hand-collected dataset to help estimate their effects. We develop a novel identification strategy, which we term Factor-Augmented Local Projection (FALP), to investigate the subtly different impacts of both monetary and macroprudential policies. Monetary policy acted on output and inflation broadly in line with consensus views today, but credit controls had markedly different effects and acted primarily to modulate bank lending.
    Keywords: Monetary policy; macroprudential policy; credit controls
    JEL: E50 G18 N14
    Date: 2016–08–19
  12. By: Groll, Dominik; Monacelli, Tommaso
    Abstract: The desirability of flexible exchange rates is a central tenet in international macroeconomics. We show that, with forward-looking staggered pricing, this result crucially depends on the monetary authority's ability to commit. Under full commitment, flexible exchange rates generally dominate a monetary union (or fixed exchange rate) regime. Under discretion, this result is overturned: a monetary union dominates flexible exchange rates. By fixing the nominal exchange rate, a benevolent monetary authority finds it welfare improving to trade off flexibility in the adjustment of the terms of trade in order to improve on its ability to manage the private sector's expectations. Thus, inertia in the terms of trade (induced by a fixed exchange rate) is a cost under commitment, whereas it is a benefit under discretion, for it acts like a commitment device.
    Keywords: monetary union,flexible exchange rates,commitment,discretion,welfare losses,nominal rigidities
    JEL: E52 F33 F41
    Date: 2016
  13. By: Sam Schulhofer-Wohl (Federal Reserve Bank of Minneapolis); Greg Kaplan (Princeton University)
    Abstract: We use scanner data to estimate inflation rates at the household level. Households' inflation rates are remarkably heterogeneous, with an interquartile range of 6.2 to 9.0 percentage points on an annual basis. Most of the heterogeneity comes not from variation in broadly defined consumption bundles but from variation in prices paid for the same types of goods -- a source of variation that previous research has not measured. The entire distribution of household inflation rates shifts in parallel with aggregate inflation. Deviations from aggregate inflation exhibit only slightly negative serial correlation within each household over time, implying that the difference between a household's price level and the aggregate price level is persistent. The large cross-sectional dispersion and low serial correlation of household-level inflation rates together mean that almost all of the variability in a household's inflation rate over time comes from variability in household-level prices relative to average prices for the same goods, not from variability in the aggregate inflation rate. We provide a characterization of the stochastic process for household inflation that can be used to calibrate models of household decisions.
    Date: 2016
  14. By: Ippei Fujiwara (Keio University and The Australian National University (E-mail:; Timothy Kam (The Australian National University (E-mail:; Takeki Sunakawa (The University of Tokyo (E-mail:
    Abstract: We provide new insight on international monetary policy cooperation using a two-country model based on Benigno and Benigno (2006). Assuming symmetry, save for the volatility of (markup) shocks, we show that an incentive feasibility problem exists between the policymakers across national borders: The country faced with a relatively more volatile markup shock has an incentive to deviate from an assumed Cooperation regime to a Non-cooperation regime. More generally, a similar result obtains if countries differ in size. This motivates our study of a history-dependent Sustainable Cooperation regime which is endogenously sustained by a cross-country, state-contingent contract between policymakers. Under the Sustainable Cooperation regime, the responses of inflation and the output gap in both countries are different from the ones under the Cooperation and Non-cooperation regimes reflecting the endogenous welfare redistribution between countries under the state- contingent contract. Such history-contingent welfare redistributions are supported by resource transfers effected through incentive-compatible variations in the terms of trade (or net exports). Such an endogenous cooperative solution may also provide a theoretical rationale for perceived occasional cooperation between national central banks in reality.
    Keywords: Monetary policy cooperation, Sustainable plans, Welfare
    JEL: E52 F41 F42
    Date: 2016–08
  15. By: Ellen E. Meade; Yoshio Nozawa; Lubomir Petrasek; Joyce Zickler
    Abstract: Print{{p}}September 24, 2015{{p}}The Effects of FOMC Communications before Policy Tightening in 1994 and 20041{{p}}Ellen E. Meade, Yoshio Nozawa, Lubomir Petrasek, and Joyce K. Zickler{{p}}The ways in which the Federal Open Market Committee (FOMC or "Committee") communicates its views on economic and financial{{p}}conditions, the economic outlook, and its policy intentions have evolved over time. Using information from publicly available historical{{p}}policy records, we describe the Committee's communications as it approached the decisions to increase the target for the federal funds{{p}}rate in February 1994 and June 2004 and we examine the effect those communications had on policy expectations in financial markets.{{p}}Tightening in 1994{{p}}At its meeting in February 1994, the FOMC voted to increase the target federal funds rate by 25 basis points to 3.25 percent (figure 1),{{p}}the first increase in the target since 1989. At that time, the Committee did not communicate actively with the public as it does today and{{p}}its communication tools were very limited. For example, statements were not issued following FOMC meetings, changes in the federal{{p}}funds rate target were not announced, and the minutes and directive for each FOMC meeting were not released until two days after the{{p}}subsequent meeting. (The directive contains the instructions regarding the implementation of monetary policy that the FOMC issues to{{p}}the Open Market Desk at the Federal Reserve Bank of New York.) The economic projections prepared by the Committee were made{{p}}publicly available only twice a year, in February and July, in the Monetary Policy Report to the Congress (MPR) and the accompanying{{p}}testimony by the Chair. Thus, in late December 1993 and early January 1994, the most up-to-date information on the Committee's{{p}}thinking was the assessment contained in the minutes of the November meeting: "In the Committee's discussion of policy for the{{p}}intermeeting period ahead, the members generally agreed that despite various indications of a pickup in economic growth, the{{p}}underlying economic situation and the outlook for inflation had not changed sufficiently to warrant an adjustment in monetary policy." The{{p}}November directive to the Open Market Desk indicated that the Committee would maintain the current degree of reserve restraint and{{p}}had a balanced, or "symmetric," view of the likely adjustment of policy over the intermeeting period, saying that either "slightly greater" or{{p}}"slightly lesser" reserve restraint might be warranted.2{{p}}Figure 1: Federal Funds Target Rate{{p}} Source: Federal Reserve Board.{{p}}Accessible version{{p}}Despite the strength of incoming economic data over the fall and winter of 1993, market expectations for the federal funds rate did not{{p}}increase significantly. Treasury yields and forward rates at short and longer-term maturities remained near the lower end of their recent{{p}}ranges (figures 2 and 3). Data from the Blue Chip survey at the time suggest that some market participants may have underestimated{{p}}the economy's momentum and resulting inflationary pressures. The Blue Chip forecast for one-year-ahead real GDP growth (figure 4){{p}}did not increase, on net, in the six months prior to the policy tightening, and expectations for consumer price inflation one-year-ahead{{p}}edged down (figure 5). Accordingly, market participants generally appeared to expect that the federal funds rate would remain{{p}}unchanged in the near future (figure 6).{{p}}Figure 2: 10-Year Par Treasury Yield{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}1 of 9 9/25/2015 11:31 AM{{p}} Source: Federal Reserve Bank of New York, staff estimations.{{p}}Accessible version{{p}}Figure 3: 1-Year Forward Rates{{p}} Source: Federal Reserve Bank of New York, staff estimations.{{p}}Accessible version{{p}}Figure 4: 1-Year-Ahead Blue Chip Real GDP Forecasts{{p}} Source: Blue Chip economic indicators.{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}2 of 9 9/25/2015 11:31 AM{{p}}Accessible version{{p}}Figure 5: 1-Year-Ahead Blue Chip CPI Inflation{{p}} Source: Blue Chip economic indicators.{{p}}Accessible version{{p}}Figure 6: 1-Year Blue Chip Federal Funds Rate Forecasts{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}3 of 9 9/25/2015 11:31 AM{{p}} Source: Blue Chip Financial Forecasts.{{p}}Accessible version{{p}}At their meeting in December 1993, FOMC participants discussed the best way to signal to financial markets and the public that a{{p}}decision to raise rates might come soon. While some spoke of wanting a tightening decision at that meeting, Chairman Greenspan{{p}}recommended that the Committee leave the target for the federal funds rate unchanged, but prepare to tighten policy early in 1994, if{{p}}necessary. He also proposed that the directive remain symmetric and not signal any bias toward tightening during the intermeeting{{p}}period, arguing that the decision to raise interest rates--a crucial step--should be taken by a vote of the Committee and not by the{{p}}Chairman in between scheduled meetings. In response to a suggestion that the Chairman use an upcoming speaking opportunity to{{p}}convey the Committee's thinking about the outlook for monetary policy, Chairman Greenspan used his testimony before the Joint{{p}}Economic Committee on the Monday before the February FOMC meeting "to emphasize that monetary policy must not overstay{{p}}accommodation" and that "the foundations of the economic expansion are looking increasingly well-entrenched." He went on to note that{{p}}historical experience suggested that "higher price inflation tends to surface rather late in the business cycle" and that "the challenge of{{p}}monetary policy is to detect such latent instabilities in time to contain them."{{p}}When the FOMC met later that week, participants generally agreed that the economy was entering 1994 with considerable forward{{p}}momentum. They saw little likelihood of further progress in reducing inflation and a "distinct risk" of higher inflation if monetary policy{{p}}remained very accommodative. These views were reflected in the projections prepared for the upcoming MPR. Against this backdrop,{{p}}participants generally favored an increase in the target for the federal funds rate. A number argued in favor of a 50-basis-point increase{{p}}in order to get out in front of inflation, but most wanted to remove the high degree of policy accommodation gradually. Committee{{p}}members expressed concerns about the possibility of a large announcement effect because it had been so long since the Committee's{{p}}last policy rate tightening in 1989. They decided to raise the target rate by 25 basis points, and for the first time, the Chairman issued a{{p}}very brief postmeeting statement announcing the change. A fuller explanation of the Committee's views came a few weeks later in the{{p}}Chairman's monetary policy testimony to the Congress.{{p}}The tightening, as well as the release of Chairman Greenspan's brief statement announcing the decision following the FOMC meeting,{{p}}were not anticipated by the public, leading to a sharp run-up in interest rates and a temporary increase in financial market volatility. As{{p}}shown in figure 7, the 10-year nominal Treasury yield increased 14 basis points on the day of the announcement, and about 200 basis{{p}}points over the next nine months. Meanwhile, the yield curve flattened as short-term rates rose even more sharply (figure 3). Several{{p}}factors may account for the sharp rise in Treasury yields. The tightening came sooner than investors had anticipated. Moreover, the{{p}}expected path of the federal funds rate over the coming year implied by futures market quotes right before the tightening--111 basis{{p}}points--substantially underestimated the subsequent tightening--300 basis points. The more-rapid pace of tightening occurred in{{p}}response to data releases over several months that repeatedly led to a marking up of expectations for economic growth and inflation{{p}}(figures 4 and 5) and, in turn, called for a higher level of yields. There was also a significant rise in uncertainty about the paths of interest{{p}}rates, as indicated by the widening of confidence intervals on the federal funds rate in figure 8.{{p}}Figure 7: 10-Year Par Treasury Yield{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}4 of 9 9/25/2015 11:31 AM{{p}} Source: Federal Reserve Bank of New York, staff estimates.{{p}}Accessible version{{p}}Figure 8: 6-Month-Ahead 90% Confidence Intervals on the Federal Funds Rate{{p}} Note: The confidence interval is estimated using Eurodollar futures options.{{p}} Source: CME, CBOT with permission from CME Group, Inc.{{p}}Accessible version{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}5 of 9 9/25/2015 11:31 AM{{p}}In contrast to interest rates, stock prices stabilized quickly after an initial negative reaction, and fully recovered their losses later in the{{p}}tightening cycle (figure 9), likely because the economy strengthened. Measures of the implied volatility of stock prices initially increased,{{p}}but reverted to normal levels after several months (figure 10). Corporate yield spreads, particularly those faced by lower-tier issuers, also{{p}}narrowed (figure 11), consistent with investors judging that the economic outlook was improving.{{p}}Figure 9: Cumulative Return of S&P 500 Stock Index{{p}} Source: Bloomberg.{{p}}Accessible version{{p}}Figure 10: Implied Volatility of S&P 500 Stock Index{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}6 of 9 9/25/2015 11:31 AM{{p}} Source: Bloomberg.{{p}}Accessible version{{p}}Figure 11: High-yield Corporate Bond Spreads{{p}} Source: Bloomberg, Federal Reserve Bank of New York, staff estimates.{{p}}Accessible version{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}7 of 9 9/25/2015 11:31 AM{{p}}Tightening in 2004{{p}}By the time of the June 2004 FOMC meeting, the Committee was routinely issuing postmeeting statements. However, the minutes and{{p}}directive for each FOMC meeting were still not released until two days after the subsequent meeting. As a result, the postmeeting{{p}}statements garnered significant attention in financial markets. The statements at the time provided an indication of the Committee's view{{p}}about risks to the outlook for economic growth and inflation--the "balance of risks." In addition, the Committee had also begun to use a{{p}}time-dependent form of forward guidance about the likely stance of policy. For instance, the August 2003 postmeeting statement stated{{p}}that policy accommodation could be maintained "for a considerable period;" in January 2004, that forward guidance was changed to{{p}}indicate that the Committee thought it could be "patient in removing" monetary policy accommodation. Thus, the Committee used a{{p}}sequence of changes in the balance of risks and forward guidance language in the months leading up to the June 2004 tightening to{{p}}signal that its assessment of the economy was evolving and that it was getting closer to raising its target for the federal funds rate.{{p}}Over the months leading up to the tightening in June 2004, financial market expectations for the timing of the first increase in and the{{p}}subsequent path for the target federal funds rate shifted noticeably in response to the receipt of economic data and FOMC{{p}}communications. During the late summer and fall of 2003, core PCE inflation had slipped to 1¼ percent and the Committee had noted in{{p}}its postmeeting statements the risk that inflation could become "undesirably low" as its "predominant concern for the foreseeable future."{{p}}By December 2003, the risks of further disinflation had come down, and the Committee stated that the risk of a fall in inflation was{{p}}"almost equal" to that of a rise. At the March 2004 meeting, the Committee decided to retain its assessment that the risks to economic{{p}}growth were "roughly equal" and the risks to inflation were "almost equal" even though some participants noted that rising energy and{{p}}commodity prices along with reports of pricing power in some sectors suggested that the risks to inflation had become more balanced,{{p}}because many Committee members continued to believe that slack in labor and output markets would keep inflation low. At the meeting,{{p}}Chairman Greenspan talked about how a move to full balance in describing the risks to inflation would be consistent with "continuing to{{p}}adjust our statements meeting by meeting and fostering a momentum that might enable us at the appropriate time to tighten with{{p}}minimum effect on rates." However, he argued that, although it was "a very close call," he wanted to "avoid creating momentum toward a{{p}}policy action if, indeed, the economic expansion was slowing down."{{p}}Financial markets read the Committee's assessment of the economy in its March statement as having a soft cast. However, by May, a{{p}}strong employment report, higher core CPI inflation, and some signs of a pickup in labor compensation led to rapidly shifting{{p}}expectations, both among policymakers and in financial markets. Near-term inflation expectations had been rising (figure 5), although{{p}}longer-run measures of inflation expectations remained stable. Bonds sold off in response to heightened prospects for policy tightening{{p}}and interest rates moved up noticeably (figures 2 and 3).{{p}}At the May 2004 meeting, policymakers revised the statement language to indicate that the risks to price stability had "moved into{{p}}balance." In addition, they replaced the forward guidance language indicating that the Committee would be "patient in removing" policy{{p}}accommodation with "the Committee believes that policy accommodation can be removed at a pace that is likely to be measured." In{{p}}discussing how to position the Committee to move at an upcoming meeting, several participants wanted to send a signal that, while the{{p}}Committee might want to start to tighten early, it intended to be gradual; a number of them specifically mentioned that they wanted to{{p}}avoid having financial markets assume that the Committee might be as aggressive as it had been during the 1994–95 tightening cycle.{{p}}Several also noted that the sentence describing their expected policy path would indicate that inflation was still relatively low and{{p}}resource slack remained, and it would therefore lend support to the expectation of a gradual pace of tightening. The use of the words{{p}}"likely to be measured" seemed to satisfy those who wanted the flexibility to move more aggressively, if needed, particularly if the{{p}}outlook for inflation worsened.{{p}}The replacement of the sentence in the statement that the Committee could be "patient" in removing policy accommodation with new{{p}}guidance that accommodation could be removed at a "measured pace" did not lead to a significant reaction in financial markets.{{p}}However, a subsequent series of economic releases showing strong gains in employment and spending caused investors to anticipate{{p}}that policy tightening would begin with an increase of 25 basis points in the target federal funds rate at the June meeting and would be{{p}}followed by similar increases at each of the remaining four meetings of the year. Measures of interest rate uncertainty (figure 8) also{{p}}moved up noticeably.{{p}}At the June 2004 meeting, most participants agreed that the Committee needed to move policy back to neutral over a period of time, but{{p}}a few indicated that they preferred removing any characterization of possible future policy actions from the postmeeting statement. Two{{p}}members expressed reservations about retaining the "measured pace" language because it was likely to imply a steady tightening path{{p}}of 25 basis points each meeting and might result in the Committee falling behind the curve as inflation rose. However, most viewed the{{p}}addition of the sentence stating that "nonetheless, the Committee will respond to changes in economic prospects as needed to fulfill its{{p}}obligation to maintain price stability" as providing scope for adjusting the path for the funds rate in response to incoming economic{{p}}information and their interpretation of its implications for economic activity and inflation.{{p}}In part because of the change in the forward guidance in May, the market reaction to the June 2004 FOMC action was much more{{p}}muted than the response to the tightening in February 1994. The 10-year Treasury yield in fact decreased 8 basis points on the day of{{p}}the initial target rate increase (figure 7), and interest rates declined further across the term structure over subsequent months (figure 3).{{p}}Standard measures of volatility and liquidity in the Treasury securities market held fairly steady at typical levels, and the confidence{{p}}interval on the federal funds rate narrowed (figure 8), indicating that interest rate uncertainty decreased. Equity prices initially declined{{p}}(figure 9), but quickly recovered amid declines in equity market volatility (figure 10). Corporate bond spreads narrowed (figure 11) and{{p}}inflation expectations stabilized (figure 5).{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}8 of 9 9/25/2015 11:31 AM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}Conclusion{{p}}The policy record shows that the FOMC's decisions about the appropriate time to begin to remove policy accommodation in February{{p}}1994 and June 2004 depended importantly on the evolution of policymakers' outlook for real economic activity, labor market conditions,{{p}}and inflation. However, at the time of the February 1994 tightening, the Committee's policy communications with the public were limited{{p}}and were not timely, and financial markets did not fully anticipate the Committee's decision. By the time of the June 2004 tightening, the{{p}}Committee's policy communications had expanded significantly--in particular, the postmeeting statements that included assessments of{{p}}the balance of risks and forward guidance were important in shaping investor expectations by signaling the Committee's assessment of{{p}}how economic conditions were evolving and its thinking about when to raise the federal funds rate target.{{p}}1. The authors thank Bill English and Steve Meyer for comments, and Edward Atkinson, Eric Horton, and Blake Phillips for research assistance. Return to text{{p}}2. The directive specified the Committee's short-term operating objective in terms of a "degree of pressure on reserve positions." It also provided the{{p}}Committee's inclination toward modifying policy over the intermeeting period--the "bias" or "tilt." Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: September 24, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: The Effects of FOMC Communications before Policy...{{p}}9 of 9 9/25/2015 11:31 AM
    Date: 2015–09–24
  16. By: Ralf R. Meisenzahl
    Abstract: March 24, 2015{{p}}The Federal Reserve's Overnight and Term Reverse Repurchase Agreement Operations in the Financial Accounts of the United States{{p}}{{p}}Ralf R. Meisenzahl1{{p}}{{p}}This note explains how the Federal Reserve's overnight and term reverse repurchase agreement (RRP) operations are reported in the Federal Reserve's Financial Accounts of the United States (formerly known as the Flow of Funds Accounts). Beginning with the March 2015 publication of the Financial Accounts, the securities repurchase transactions of the monetary authority, which include monetary authority repurchase agreements (RPs) and RRPs, will show, as separate line items (in tables F.109, link, and L.109, link), (1) RRP operations that have been conducted as part of the Federal Reserve's Overnight Reverse Repurchase Agreement Operational Exercise since September 2013, combined with term RRP operations that were first conducted in December 2014, and (2) other RPs and RRPs.2 The overnight and term RRP operations will be also appear as memo items on the federal funds and security repurchase agreements instrument tables (F.207, link, and L.207, link). Further details are provided below.{{p}}{{p}}Background on the Federal Reserve's overnight and term RRP operations{{p}}Federal Reserve RPs and RRPs are conducted by the Open Market Trading Desk (the "Desk") at the Federal Reserve Bank of New York. In an RP transaction, the Desk purchases a U.S. government security from an eligible counterparty that agrees to repurchase the same security at a specified price at a specific time in the future. An RRP transaction is the opposite of an RP: The Desk sells a U.S. government security and agrees to repurchase it later. For an overnight RP / RRP, the time between the purchase and repurchase is one business day, while for term RPs / RRPs, the interval can be as many as 65 business days.{{p}}{{p}}The Federal Open Market Committee (FOMC) first authorized the Desk to conduct a series of fixed-rate overnight RRP operations involving U.S. Government securities, including agency securities, in September 2013.3 The FOMC has indicated that it plans to use an overnight RRP facility to help control the federal funds rate during the monetary policy normalization process.4{{p}}{{p}}The overnight RRP operational exercise started on September 23, 2013. Take-up was initially limited to $0.5 billion per counterparty, although this limit was raised to $1 billion a few days later. The limit was subsequently raised in a series of steps, and reached $30 billion per counterparty by late September 2014, when an aggregate limit of $300 billion was also imposed on each overnight RRP operation. The total number of eligible RRP counterparties has increased over the duration of the exercise from 140 in September 2013 to 164 in March 2015.5{{p}}{{p}}In October 2014, the FOMC authorized a series of term RRP operations with an aggregate limit of $300 billion to begin in December 2014 and mature in early January 2015.6 In January 2015, the FOMC authorized the Desk to conduct $200 billion of term RRPs over the March 2015 quarter-end.7 Figure 1 shows combined quarter-end take-up of overnight and term RRP (small amounts of term RRPs that were conducted in testing prior to the fourth quarter of 2014 are not included).{{p}}Figure 1: Quarter-end Take-up in Federal Reserve RRP operations{{p}}Figure 1: Quarter-end Take-up in Federal Reserve RRP operations. See accessible link for data.{{p}}{{p}}{{p}} Source: Federal Reserve Bank of New York:{{p}}{{p}}Accessible version{{p}}{{p}}Reporting overnight and term RRP operations in the Financial Accounts{{p}}Figure 2 shows the effects of an RRP transaction on the monetary authority's and an RRP counterparty's balance sheets. The asset side of the monetary authority's balance sheet is unaffected; it continues to show the securities that have been sold temporarily under the RRP operations. The effect of the RRP is to shift the composition of the monetary authority's liabilities. Its RP ("repo") liabilities expand by $100 (recall that RPs are the opposite of RRPs). At the same time, its reserve liabilities--deposits held by banks and other depository institutions in their accounts at the Federal Reserve--decrease by $100. That is, when a counterparty lends cash to the Federal Reserve through RRPs, the Federal Reserve receives the cash by debiting the reserve account of the bank that clears the counterparty's trade. The RRP operational exercise therefore does not change the overall size of the monetary authority's balance sheet.{{p}}Figure 2: Effects of Reverse Repurchase Agreement on Balance Sheets{{p}}Figure 2: Effects of Reverse Repurchase Agreement on Balance Sheets. See accessible link for data.{{p}}{{p}}Accessible version{{p}}{{p}}Starting with the March 2015 publication of the Financial Accounts of the United States, the balance sheet of the monetary authority (table L.109, link) includes two memo items under securities repurchase agreements. These items are (1) the combined amount of Federal Reserve liabilities in the overnight and term RRP operations as of the last day of each quarter, as shown in figure 1, and (2) Federal Reserve net liabilities under other RPs, which mostly have foreign official and international accounts as counterparties.{{p}}{{p}}Eligible counterparties for Federal Reserve's overnight and term RRP operations include four types of financial institutions: primary dealers, banks (domestic and foreign banking offices in the United States), government-sponsored enterprises (GSE), and money market mutual funds. Figure 3 shows that the dominant type of counterparty for the quarter-end overnight RRP operations has been money market mutual funds.8{{p}}Figure 3: Quarter-end RRP Take-up by Type of Financial Institution{{p}}Figure 3: Quarter-end RRP Take-up by Type of Financial Institution. See accessible link for data.{{p}}{{p}}{{p}} Source: Federal Reserve Bank of New York:{{p}}{{p}}Accessible version{{p}}{{p}}In addition to the monetary authority sector table, The Financial Accounts of the United States also reports volumes of overnight and term RRP operations on the instrument table for Federal Funds and Security Repurchase Agreements (table L.207, link). On this table, the RRP operations appear as a memo item showing the monetary authority's RRP liabilities and money market funds' and other financial institutions' RRP assets.9{{p}}{{p}}References{{p}}Frost, Josh, Lorie Logan, Antoine Martin, Patrick McCabe, Fabio Natalucci, and Julie Remache (2015). "Overnight RRP Operations as a Monetary Policy Tool: Some Design Considerations (PDF)," Finance and Economics Discussion Series 2015-010. Board of Governors of the Federal Reserve System (U.S.).{{p}}{{p}}1. I would like to thank Brian Bonis, Jane Ihrig, Patrick McCabe, John McGowan, Maria Perozek, William Riordan and Paul Smith for helpful comments. Return to text{{p}}{{p}}2. Other RPs and RRPs include those conducted with foreign official and international counterparties. Weekly Federal Reserve's balance sheet data, including all RPs and RRPs, are reported in the Federal Reverse's H.4.1 release ( Return to text{{p}}{{p}}3. The Minutes from the September 2013 FOMC meeting, which report the first authorization of overnight RRP operations, can be found at In December 2014, the FOMC authorized overnight RRP operations until January 29, 2016 ( For a detailed description of the RRP operational exercise, see Frost et al (2015). Return to text{{p}}{{p}}4. See the FOMC's Policy Normalization Principles and Plans, published in September 2014, which can be found at Return to text{{p}}{{p}}5. Lists of currently eligible counterparties is published by the Federal Reserve Bank of New York. The list of primary dealers serving as trading counterparties for the implementation of monetary policy can be found at The expanded list of counterparties that are also eligible for participation in RRP operations can be found at Return to text{{p}}{{p}}6. The FOMC authorization of the term RRP operations is included in the Minutes from the October 2014 FOMC meeting, available at Term RRP amounts can be retrieved from Return to text{{p}}{{p}}7. See Return to text{{p}}{{p}}8. The Federal Reserve Bank of New York publishes the composition of the daily overnight RRP take-up data with a three month delay at Return to text{{p}}{{p}}9. Other include primary dealers, banks (domestic and foreign banking offices in the United States), and government-sponsored enterprises. Return to text{{p}}{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Search Working Papers{{p}}{{p}}Last update: March 24, 2015
    Date: 2015–03–24
  17. By: Alan K. Detmeister; Daeus Jorento; Emily Massaro; Ekaterina V. Peneva
    Abstract: June 8, 2015{{p}}Did the Fed's Announcement of an Inflation Objective Influence Expectations?{{p}}Alan Detmeister, Daeus Jorento, Emily Massaro, and Ekaterina Peneva{{p}}Economic theory suggests that inflation expectations are a key determinant of actual inflation. In particular, without shocks from labor{{p}}markets, movements in oil prices, exchange rates or idiosyncratic factors, prices should change at an average pace consistent with{{p}}expectations. Indeed, empirical work attests to their importance: conditioning on long-run inflation expectations from surveys improves{{p}}the accuracy of inflation forecasts.1 Thus, an important question is whether, and how, the Federal Open Market Committee (FOMC) can{{p}}influence inflation expectations.{{p}}Figure 1 shows three survey measures of long-run expected inflation: the expected rate of price change during the next five to ten years{{p}}from the University of Michigan's Survey of Consumers in blue; the expected average Consumer Price Index (CPI) inflation during the{{p}}next ten years from the Philadelphia Fed's Survey of Professional Forecasters, or SPF, in red; and expected Personal Consumption{{p}}Expenditure (PCE) price inflation during the next ten years from the SPF in black.{{p}}Figure 1: Survey Measures of Long-Run Expected Inflation, 1992-Present{{p}}Note. Median responses.{{p}} Sources: University of Michigan, Surveys of Consumers ; Federal Reserve Bank of Philadelphia, Survey of Professional Forecaster (SPF) .{{p}}Accessible version{{p}}As can be seen from the chart, these survey measures have moved little since the late 1990s despite the recent deep recession,{{p}}significant swings in commodity prices, and unprecedented monetary policy actions. The remarkable stability of these measures,{{p}}however, makes it very difficult to figure out what drives them using standard time series regression techniques. For this reason, we look{{p}}at how inflation expectations have changed following a single event--the FOMC's announcement of a 2 percent longer-run objective for{{p}}PCE price inflation in January 2012.{{p}}Starting with the expectations of professional forecasters, in the four quarters following FOMC's announcement, expected PCE price{{p}}inflation during the next 10 years--the solid black line in figure 2--moved from about 1/4 percentage point above the FOMC's objective{{p}}down to the 2 percent objective. In the two years since coming into line with the FOMC's objective, these expectations have remained at{{p}}the 2 percent level.{{p}}Figure 2: Long-Run Expected Inflation from the SPF, 2011-Present{{p}} FRB: FEDS Notes: Did the Fed's Announcement of an Inflation Objectiv...{{p}}1 of 5 6/9/2015 11:07 AM{{p}}Note. Median responses.{{p}} Sources: Federal Reserve Bank of Philadelphia, Survey of Professional Forecaster (SPF) .{{p}}Accessible version{{p}}Arguably, since monetary policy has its greatest influence on inflation in the longer run, removing the initial five years from the ten year{{p}}sample provides a clearer signal of the inflation rate that professional forecasters think is consistent with monetary policy. The black{{p}}dashed line in figure 2 displays professional forecasters' expectations of PCE inflation in years six through ten. These expectations{{p}}trended down more gradually following the announcement, and have just recently reached the FOMC's 2 percent inflation objective. 2 ,3{{p}}Turning to the CPI, expected inflation during the next 10 years--the solid red line--also declined since the announcement, though fairly{{p}}gradually. Given that inflation as measured by the CPI has been slightly higher, on average, than inflation measured by PCE prices, it is{{p}}no surprise that CPI expectations remain above 2 percent. In contrast to the downward movement in the other measures, expected CPI{{p}}inflation in years six through ten--the red dashed line--has shown little net change.{{p}}Taken at face value, the movement in CPI and PCE price inflation expectations provide a little support for the idea that the FOMC's{{p}}announced inflation objective did influence the inflation expectations of professional forecasters, who probably follow FOMC statements{{p}}fairly closely.4{{p}}On the other hand, there is little evidence that households were influenced by the FOMC's announcement. As shown by the blue line in{{p}}figure 1, expected inflation during the next five to ten years from the Michigan survey continued to move sideways after early 2012. It is{{p}}worth noting that Michigan survey does not specify a particular price measure, and the time horizon is less explicit than in the SPF.{{p}}Examining individual responses from the SPF and Michigan survey provides further support for the idea that professional forecasters{{p}}were influenced by the FOMC's announcement, but households were not. As can be seen by comparing the blue and orange bars in{{p}}figure 3, the expectations of professional forecasters of PCE price inflation during the next 10 years became more concentrated around{{p}}2 percent in the three years following the FOMC's January 2012 announcement compared to the three years prior to the announcement.{{p}}Both professional forecasters with very low inflation expectations and very high inflation expectations have become less common, and{{p}}the standard deviation of across participants' responses declined from an average of 0.62 percentage points in the surveys during three{{p}}years prior to the announcement to 0.40 percentage points in the surveys during the three years since. 5 By contrast, the distribution of{{p}}long-run Michigan inflation expectations--figure 4--reveals little if any change in consumer inflation expectations following the{{p}}announcement of a longer-run inflation objective.{{p}}Figure 3: Distribution of SPF Expectations for 10-year PCE Inflation{{p}} FRB: FEDS Notes: Did the Fed's Announcement of an Inflation Objectiv...{{p}}2 of 5 6/9/2015 11:07 AM{{p}} Sources: Authors' calculations using data from Federal Reserve Bank of Philadelphia, Survey of Professional Forecaster (SPF){{p}}Accessible version{{p}}Figure 4: Distribution of Michigan Expectations for 5- to 10-Year Inflation{{p}} Sources: Authors' calculations using data from University of Michigan, Surveys of Consumers .{{p}}Accessible version{{p}}As yet another way of describing how longer-term inflation expectations have changed over time, the next two figures plot annual time{{p}}series of the fraction of SPF and Michigan survey respondents with a long-term inflation expectation of 2 percent. The average share of{{p}}professional forecasters who expected exactly 2 percent PCE inflation during the next 10 years--figure 5--rose noticeably from the three{{p}}years prior to the three years following the FOMC's announcement, though the share after the announcement is no higher than it was in{{p}}2007 and 2008. This increase in the share is statistically significant.6 At the same time, the average share of consumers in the Michigan{{p}}survey who expected 2 percent inflation over the longer term--figure 6--changed little, with the difference being neither economically nor{{p}}statistically significant.7{{p}}Figure 5: Share of SPF Respondents Who Expect Exactly 2 Percent PCE Inflation Over the Next 10 Years{{p}} FRB: FEDS Notes: Did the Fed's Announcement of an Inflation Objectiv...{{p}}3 of 5 6/9/2015 11:07 AM{{p}} Sources: Authors' calculations using data from Federal Reserve Bank of Philadelphia, Survey of Professional Forecaster (SPF) .{{p}}Accessible version{{p}}Figure 6: Share of Michigan Survey Respondents Who Expect 2 Percent Inflation Over the Next 5 to 10 Years{{p}} Sources: Authors' calculations using data from University of Michigan, Surveys of Consumers .{{p}}Accessible version{{p}}To conclude, the data suggest that the FOMC's announcement of an explicit inflation objective had some effect on professional{{p}}forecasters' long-run inflation expectations, but not on households' expectations. Admittedly, inflation expectations of professional{{p}}forecasters did not immediately jump to the FOMC's objective, so it is not clear just how much of change in professional forecasters'{{p}}expectations can be attributed to the FOMC's announcement versus other factors, such as a general reduction in uncertainty as{{p}}economic conditions improved and actual inflation remained moderate. That said, given that the decline in the SPF inflation{{p}}expectations, the tightening of the distribution, and the increase in the share expecting exactly 2 percent inflation started around the time{{p}}of the announcement, it is likely the FOMC's announcement influenced the views of professional forecasters. We cannot say the{{p}}FOMC's announcement had the same influence on households.{{p}} References:{{p}}Binder, Carola, "Fed Speak on Main Street," working paper (2014).{{p}}Clark, Todd E., and Taeyoung Doh, "Evaluating Alternative Models of Trend Inflation," International Journal of Forecasting, Volume 30,{{p}}Issue 3 (2014), 426-448.{{p}} FRB: FEDS Notes: Did the Fed's Announcement of an Inflation Objectiv...{{p}}4 of 5 6/9/2015 11:07 AM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}Faust, Jon, and Jonathan H. Wright, "Forecasting Inflation," in G. Elliott and A. Timmerman (Eds.) Handbook of Economic Forecasting,{{p}}volume 2A. Amsterdam: North Holland (2013).{{p}}Nechio, Fernanda, "Have Long-term Inflation Expectations Declined?" Federal Reserve Bank of San Francisco Economic Letter, no.{{p}}2015-11 (2015).{{p}}Zaman, Saeed, "Improving Inflation Forecasts in the Medium to Long Term," Federal Reserve Bank of Cleveland Economic{{p}}Commentary, no. 2013-16. November 16 (2013){{p}}1. This may be because long-run inflation expectations proxy quite well for the trend in inflation. Including a measure of the trend improves the quality of{{p}}inflation predictions. See, among others, Faust and Wright (2013), Clark and Doh (2014), and Zaman (2013). Return to text{{p}}2. The figure starts in 2011, the time when the SPF implemented a check to confirm that respondents' expectations for inflation 6 to 10 years ahead were{{p}}consistent with their answers about expected inflation during the next 5 and the next 10 years. In order to use longer sample period we focus on SPF{{p}}expectations for 10-year PCE inflation in the rest of the analysis. Return to text{{p}}3. Nechio (2015) finds that the decline in expected PCE inflation in years six through ten in the SPF survey is primarily driven by revised expectations from{{p}}forecasters who overestimated inflation in the aftermath of the Great Recession. Return to text{{p}}4. In addition, in the second quarter of 2012, the SPF panelists were explicitly told that on January 25th, 2012 the FOMC had reached a broad agreement on{{p}}some principles regarding its long-run goals, including that "The committee judges that inflation at the rate of 2 percent, as measured by the annual change in{{p}}the price index for personal consumption expenditures, is most consistent over the longer run with the Federal Reserve's statutory mandate" In the same{{p}}quarter, the Philadelphia Fed's survey of Professional Forecasters included a special question which asked the panelists "...whether their long-run forecast for{{p}}inflation in the price index for personal consumption expenditures (PCE) differs in an economically meaningful way from the FOMC's longer-run goal for{{p}}inflation of 2 percent." About three-quarters of the panelists who answered the special question indicated that their long-run forecasts for PCE inflation did not{{p}}differ from the FOMC's goal in an economically meaningful way. The rest thought that inflation in the long run will exceed 2 percent and the FOMC will not{{p}}achieve its goal. Return to text{{p}}5. Statistical tests strongly reject the equality of the variances in the responses of professional forecasters across the two periods. A similar narrowing of the{{p}}distribution is observed for PCE inflation in years 6 to 10 (not shown). Relatedly, the differences between the upper and lower quartiles for all SPF-based{{p}}measures of longer-run expected inflation (not shown) have declined noticeably since 2012, with the interquartile ranges for the 10-year and six-to-ten-year{{p}}forward measures of expected PCE inflation narrowing to historically low levels. Return to text{{p}}6. We tested for difference in means between the two three-year periods, prior and following the FOMC's inflation objective announcement. The errors were{{p}}corrected for heteroskedasticity and autocorrelations using a Newey-West procedure with 1 lag (the results were not sensitive to using 4 lags). Return to text{{p}}7. Based on analysis of Michigan survey micro data, Binder (2014) concludes that the announcement of an inflation objective has not reached the general{{p}}public and inflation expectations of consumers are weakly anchored. Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: June 8, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: Did the Fed's Announcement of an Inflation Objectiv...{{p}}5 of 5 6/9/2015 11:07 AM
    Date: 2015–06–08
  18. By: Denis Gorea; Oleksiy Kryvtsov; Tamon Takamura
    Abstract: This note examines the merits of monetary policy adjustments in response to financial stability concerns, taking into account changes in the state of knowledge since the renewal of the inflation-targeting agreement in 2011. A key financial system vulnerability in Canada is elevated household indebtedness: as more and more households are nearing their debt-capacity limits, the likelihood and severity of a large negative correction in housing markets are also increasing. Adjusting the path of policy rates can be effective in reducing the buildup of household debt and the likelihood of a house price correction over the medium term. Such adjustments can also generate a fall in inflation and in output over the short term compared with the case without a policy-rate adjustment. Overall, the estimated benefits of a leaning adjustment tend to be smaller than its social losses, since its impact on the buildup of vulnerabilities is modest and the reduction in the incidence of house price corrections or financial crises is limited.
    Keywords: Financial stability, Monetary policy framework
    JEL: E0 E44 E52 E58 G18
    Date: 2016
  19. By: Peter Hördahl; Jhuvesh Sobrun; Philip Turner
    Abstract: International linkages between interest rates in different currencies are strong, and ultra-low rates have become a global phenomenon. This paper compares how interest rates in advanced economies and in emerging economies are conditioned by two global benchmarks - the Federal funds rate at the short end and the "world" real interest rate at the long end. Real equilibrium policy rates (the natural rate) have fallen in many countries, and short-term rates worldwide have been further depressed by many years of the US policy rate close to zero. Nevertheless, changes in the Federal funds rate have less effect on longer-term rates, and thus on financing conditions, than is often supposed. The decline in the world long-term rate since 2008 has been driven almost entirely by a fall in the world term premium (negative in nominal terms since mid-2014). The world short-term rate expected over the long run has fallen only modestly over the past seven years or so, and is now just over 2% (compared with around 4% pre-Lehman).
    Keywords: bond markets, financial globalization, natural rate of interest, term premium and shadow policy rate
    Date: 2016–08
  20. By: Diego Perez (New York University); Pablo Ottonello (University of Michigan)
    Abstract: The currency composition of sovereign external debt in emerging market economies is tilted towards foreign currency and the share of debt denominated in local currency is highly pro-cyclical. We study these facts through the lens of a quantitative model of optimal currency-composition of sovereign debt when the government lacks commitment regarding monetary policy. High levels of debt in local currency give rise to incentives to dilute debt repayment through nominal currency depreciation. Governments tilt the currency-composition of debt towards foreign currency to avoid the inflationary costs associated with currency depreciation. This is done at the expense of foregoing the hedging properties of debt in local currency. The cyclicality of the currency-composition of sovereign debt responds to the cyclical properties of the benefits associated to debt dilution, which are higher in recessions. Inflation-linked bonds do not eliminate the time inconsistency problem of monetary policy.
    Date: 2016
  21. By: Francesco Bianchi; Martin Lettau; Sydney C. Ludvigson
    Abstract: This paper presents evidence of infrequent shifts, or "breaks," in the mean of the consumption-wealth variable cay_{t}, an asset market valuation ratio driven by fluctuations in stock market wealth relative to economic fundamentals. Conventional estimates of cay_{t}, which presume a constant mean, display increasing persistence over the sample. We introduce a Markov-switching version of cay_{t} that adjusts for infrequent shifts in its mean. The Markov-switching cay_{t}, denoted cay_{t}^{MS}, is less persistent and has superior forecasting power for excess stock market returns compared to the conventional estimate. Evidence from a Markov-switching VAR shows that these low frequency swings in post-war asset valuation are strongly associated with low frequency swings in the long-run expected value of the Federal Reserve's primary policy rate, with low expected values for the real federal funds rate associated with high asset valuations, and vice versa. By contrast, there is no evidence that the infrequent shifts to high asset valuations and low policy rates are associated with higher expected economic growth or lower economic uncertainty; indeed the opposite is true.
    JEL: E02 E4 E52 G12
    Date: 2016–08
  22. By: Marcelo Rezende; Mary-Frances Styczynski
    Abstract: June 19, 2015{{p}}The Effects of Liquidity Regulation on Participation in the Term Deposit Facility{{p}}Marcelo Rezende and Mary-Frances Styczynski1{{p}}This note studies the effects of liquidity regulation on the participation of commercial banks in Term Deposit Facility (TDF) operations.{{p}}We find that banks subject to the Liquidity Coverage Ratio (LCR) submit tenders relatively more often than banks not subject to it in{{p}}operations that allow banks to withdraw funds prior to the maturity date. Given that term deposits must have an early withdrawal feature{{p}}(EWF) in order to qualify as high-quality liquid assets (HQLA) under LCR requirements, this result suggests that the LCR influences{{p}}banks' participation in TDF operations.{{p}}Background on the TDF and LCR Regulation{{p}}The TDF is a tool created by the Federal Reserve that could be used to drain reserves as part of the process of monetary policy{{p}}normalization. The Federal Reserve offers term deposits through the TDF to depository institutions that are eligible to receive interest{{p}}from Federal Reserve Banks.2 The funds placed in a term deposit are deducted from the institution's reserve account for the life of the{{p}}term deposit and thus cannot be used to satisfy an institution's reserve balance requirement.3 From a monetary policy perspective, the{{p}}removal of balances from institutions' reserve accounts drains reserves from the banking system, enhancing policymakers' ability to{{p}}control money market rates. From the institutions' perspective, term deposits are an asset that they hold with the central bank that earns{{p}}an interest rate that exceeds the interest rate paid on excess reserves.{{p}}As part of the tests of the TDF, the Federal Reserve has changed many characteristics of the term deposits offered. Characteristics that{{p}}have varied across operations include the operation format (offering a fixed or floating interest rate), the maturity, the interest rate, the{{p}}time between the operation and its settlement, and the maximum tender amount. In addition, while past operations did not allow{{p}}institutions to withdraw funds prior to maturity, all operations from October 2014 to the present include an EWF, subject to a pecuniary{{p}}penalty. The penalty for withdrawing a term deposit early involves the forfeiture of all interest and an annual penalty rate of 0.75 percent{{p}}applied to the term deposit principal amount.{{p}}The EWF should increase demand for term deposits among large institutions in particular because, like excess balances, term deposits{{p}}with an EWF help these institutions to meet LCR requirements. The LCR is the ratio of an institution's HQLA amount to its projected net{{p}}cash outflows over a 30-day period. The EWF qualified TDF funds as Level 1 HQLA in LCR calculations, and thus made TDF an{{p}}alternative to holding other Level 1 HQLA assets such as excess reserves and U.S. Treasury securities.4{{p}}The LCR applies to large institutions and will be implemented gradually over the next two years. The standard LCR applies to all banking{{p}}organizations with $250 billion or more in total consolidated assets or $10 billion or more in on-balance-sheet foreign exposures and to{{p}}these banking organizations' subsidiary depository institutions with assets of $10 billion or more. The modified LCR, a less stringent{{p}}version of the LCR, applies to bank holding companies and savings and loan holding companies that do not meet these thresholds but{{p}}have $50 billion or more in total assets. Institutions subject to the standard version must have an LCR of at least 80, 90, and 100 percent{{p}}by January 2015, 2016, and 2017, respectively, while institutions subject to the modified version must have an LCR of at least 90 and{{p}}100 percent by January 2016 and 2017, respectively. Thus, in 2014, when most of the TDF operations in our sample were conducted,{{p}}the LCR did not apply to any institutions; and in 2015, when the other operations in our sample were conducted, it applied only to{{p}}standard LCR institutions. However, LCR institutions should be relatively more interested in gaining familiarity with the TDF before the{{p}}LCR applies to them. In this note, we investigate whether the EWF did indeed cause an increase in the participation of LCR banks that{{p}}exceeds the change in participation by non-LCR banks.{{p}}Data{{p}}We use a panel in which each observation is a commercial bank-TDF operation pair. The panel is composed of the 3,924 domestically{{p}}chartered commercial banks that were eligible to participate in all 15 TDF operations offering maximum award amounts greater than or{{p}}equal to $10 billion. These operations were conducted between May 19, 2014, and February 19, 2015.5 Domestic banks include U.S.{{p}}chartered commercial banks and exclude U.S. branches or agencies of foreign banking organizations.{{p}}We restrict the data to domestic institutions because their participation in TDF operations should be more sensitive to U.S. liquidity{{p}}regulation. We also restrict the data to commercial banks, instead of all depository institutions that can participate in TDF operations,{{p}}because HQLA data are available only for commercial banks.6 In addition, we restrict the data to operations with a maximum award{{p}}amount of at least $10 billion to eliminate observations from operations in which large banks possibly did not submit tenders because{{p}}they considered the maximum award amounts too low.{{p}}For each bank-operation pair, the data include the dollar amounts of the bank's total excess reserves in the most recent week before the{{p}}operation, which are obtained from internal Federal Reserve databases. The data also include the total assets and HQLA from the most{{p}} FRB: FEDS Notes: The Effects of Liquidity Regulation on Participation i...{{p}}1 of 5 6/22/2015 1:36 PM{{p}}recent quarter before the operation, which are obtained from quarterly reports of condition and income.7{{p}}Effects of the LCR on Participation in TDF Operations{{p}}In this section, we investigate whether the LCR affects the decision of banks to submit tenders in TDF operations. We start analyzing{{p}}average participation rates of banks depending on whether they are subject to the LCR and whether the operations have an EWF. We{{p}}use a simple indicator of whether a bank is subject to the LCR. This indicator is equal to 1 if the bank holds $50 billion or more of total{{p}}assets and equal to 0 otherwise. Thus, this indicator does not account for cases in which a bank with assets between $10 billion and $50{{p}}billion is subject to the LCR because it is affiliated with a banking organization with $250 billion or more in total consolidated assets or{{p}}$10 billion or more in on-balance-sheet foreign exposures. However, the results do not change much if the indicator also accounts for{{p}}these cases because of the small number of such banks.{{p}}As shown in table 1, banks subject to the LCR submit tenders relatively more often than banks not subject to it in operations with an{{p}}EWF. Non-LCR banks participate in less than 1 percent of operations with and without an EWF (panels 1.a and 1.b), while LCR banks{{p}}participate in 21 percent and 36 percent of operations without and with an EWF (panels 2.a and 2.b), respectively, suggesting that the{{p}}LCR causes a larger response to an EWF.8 However, these summary statistics do not account for other characteristics of bank-operation{{p}}pairs that may determine whether a bank is subject to the LCR, an operation has an EWF, or a bank participates in an operation. For this{{p}}purpose, we next describe an empirical strategy that accounts for those characteristics and allows us to properly investigate whether{{p}}TDF participation depends on whether a bank is subject to the LCR and whether the operation has an EWF.{{p}}Table 1: Summary Statistics of Banks and TDF Tenders{{p}}Variable Obs. Mean Std. Dev. Min. Max.{{p}}1. Banks with less than $50 billion in assets{{p}}a. Operations without an EWF{{p}}Submitted a TDF tender (percentage points) 19,394 0.48{{p}}TDF tender amount 19,394 1 40 0 1,740{{p}}Assets 19,394 792 2,583 4 46,229{{p}}Excess reserves 19,394 33 271 0 12,393{{p}}HQLA 19,394 95 496 0 15,323{{p}}b. Operations with an EWF{{p}}Submitted a TDF tender (percentage points) 38,882 0.57{{p}}TDF tender amount 38,882 2 55 0 5,700{{p}}Assets 38,882 828 2,705 4 48,353{{p}}Excess reserves 38,882 37 285 0 13,674{{p}}HQLA 38,882 97 577 0 20,936{{p}}2. Banks with $50 billion or more in assets{{p}}a. Operations without an EWF{{p}}Submitted a TDF tender (percentage points) 165 21.21{{p}}TDF tender amount 165 978 2,496 0 10,000{{p}}Assets 165 310,742 475,870 60,464 2,002,047{{p}}Excess reserves 165 27,173 56,987 0 274,281{{p}}HQLA 165 70,849 120,577 4,928 546,944{{p}}b. Operations with an EWF{{p}}Submitted a TDF tender (percentage points) 330 36.36{{p}}TDF tender amount 330 3,280 6,361 0 20,000{{p}}Assets 330 323,036 493,036 53,547 2,074,952{{p}}Excess reserves 330 30,087 65,580 0 390,977{{p}}HQLA 330 81,017 132,900 3,921 656,176{{p}} Note: Each observation is a bank-operation pair. All variables are measured in millions of U.S. dollars unless stated otherwise.{{p}}Figure 1 motivates our empirical strategy. In this figure, the horizontal axis measures bank assets and the vertical axis measures{{p}} FRB: FEDS Notes: The Effects of Liquidity Regulation on Participation i...{{p}}2 of 5 6/22/2015 1:36 PM{{p}}participation rates. We divide bank-operation pairs in intervals of total assets equal to 0.2ln(7) and depending on whether the operation{{p}}has an EWF.9 Then, each point in the chart represents the average participation rate of all bank-operation pairs ij in each interval of{{p}}assets and EWF status. The curves in the chart are third-order polynomials estimated separately for observations on each side of the{{p}}$50 billion threshold and for operations with and without an EWF.10 The solid and the dashed curves show the polynomials estimated{{p}}with data from operations with and without an EWF, respectively. The curves show that participation rates jump at the threshold in{{p}}operations with and without an EWF. However, the jump is larger for banks just to the right of the threshold than for banks just to the left,{{p}}which is consistent with a positive effect of LCR coverage on TDF participation. Based on this evidence, we construct a regression{{p}}discontinuity research design to break the endogeneity between TDF participation and LCR coverage.{{p}}Figure 1: Participation of Banks in TDF Operations{{p}} Note: This figure shows the average participation rates of commercial banks in TDF operations with a maximum award amount of at least $10 billion. The{{p}}horizontal axis measures bank assets and the vertical axis measures participation rates. We divide bank-operation pairs in intervals of total assets equal to{{p}}0.2ln(7) and depending on whether the operation has an EWF. Then, each point in the chart represents the average participation rate of all bank-operation{{p}}pairs ij in each interval of assets and EWF status. The curves in the chart are third-order polynomials estimated separately for observations on each side of the{{p}}$50 billion threshold and for operations with and without an EWF. The solid and the dashed curves show the polynomials estimated with data from operations{{p}}with and without an EWF, respectively. To estimate these polynomials we restrict the sample to observations from banks with assets between $2 billion and{{p}}$1,250 billion. The chart does not show observations from banks with assets above $350 billion to maintain the information on participation confidential.{{p}}Accessible version{{p}}We estimate the following equation:{{p}}Yij = ΒI( Aij ≥ 50bi ) × EWFj + ΘXij + f ( ln(Aij ) − ln ( 50bi ) ) + νi + φj + εij, (1){{p}}where Yij is a dummy variable that is equal to 1 if bank i submitted a tender in operation j and equal to 0 otherwise. I( Aij ≥ 50bi ) is the{{p}}indicator function of whether the assets of bank i at the time of operation j were larger than or equal to $50 billion, or, equivalently, the{{p}}indicator of whether this bank was subject to the LCR during this operation. EWFj is the indicator of whether the operation has an EWF{{p}}and thus I( Aij ≥ 50bi ) × EWFj is the interaction term between the LCR and the EWF, which we use to identify the effect of the LCR on{{p}}the demand for TDF deposits. This effect is measured by the coefficient Β. Θ is a vector of coefficients and Xij is a vector of bank{{p}}characteristics, namely the ratio of excess reserves to total assets and the ratio of HQLA to total assets, which may vary across{{p}}operations. νi is a bank random effect, φj is an operation fixed effect, and εij is a bank-operation-specific unobservable error. νi and εij{{p}}have independent normal distributions with a mean of zero and variance σν{{p}}2 and σε{{p}}2. f (.) is a flexible function of assets normalized by{{p}}the LCR threshold of $50 billion. In the results presented in table 2, we assume that f (.) is a third-order polynomial that is different in{{p}}each side of the threshold, but our results are robust to changes in its functional form.11{{p}}Column 1 of table 2 shows results using all observations. The 0.15 coefficient of the interaction term between the LCR and the EWF is{{p}}statistically significant. This coefficient implies that the odds that a bank covered by the LCR submits a tender increases 15 percentage{{p}} FRB: FEDS Notes: The Effects of Liquidity Regulation on Participation i...{{p}}3 of 5 6/22/2015 1:36 PM{{p}}points more than the odds that a bank not covered by the LCR submits a tender if the operation includes an EWF.{{p}}Table 2: Measuring the Effect of LCR on TDF Participation{{p}}(1) (2){{p}}Variable All Commercial Banks Commercial Banks with Assets between $2 billion and $1,250 billion{{p}}EWF × LCR indicator 0.15** 0.12*{{p}}(0.06) (0.06){{p}}Excess reserves ratio 0.02 0.29{{p}}(0.02) (0.19){{p}}HQLA ratio 0.02 0.23{{p}}(0.01) (0.14){{p}}Observations 58,771 4,374{{p}}Banks 3,924 302{{p}}R-squared 0.15 0.17{{p}} Note: This table shows estimates of equation (1). The dependent variable is equal to 1 if bank i participated in operation j and is equal to 0 otherwise. The LCR{{p}}indicator is equal to 1 if bank i's total assets are equal to $50 billion or more and equal to 0 otherwise. Both specifications include third-order polynomials of the{{p}}natural logarithim of the difference between bank i's total assets and $50 billion that are different on each side of the $50 billion threshold. In addition, both{{p}}specifications include operation fixed effects. Standard errors are clustered at the bank level.{{p}}* and ** denote significant at the 5 and 1 percent levels, respectively. Return to text.{{p}}In regression discontinuity designs, it is standard practice to restrict the sample to observations close to the discontinuity in the{{p}}treatment, which in the case of the LCR is at $50 billion, because observations away from the discontinuity should have no weight in the{{p}}estimation of treatment effects (Lee and Lemieux, 2010). Thus, in column 2, we restrict the sample to observations with the natural{{p}}logarithm of assets at most ln(25) lower or higher than ln(50 billion)--that is, with assets between $2 billion and $1,250 billion. This{{p}}reduces the number of observations and of banks to less than a tenth of the full sample numbers--that is, from 58,771 to 4,374{{p}}observations and from 3,924 to 302 banks. Despite this restriction in the sample, the results in column 2 confirm the finding in column 1{{p}}that the LCR affects the demand for TDF deposits, but the magnitude of the effect is smaller: The 0.12 coefficient is statistically{{p}}significant and implies that the odds that a bank covered by the LCR submits a tender increases 12 percentage points more than the{{p}}odds that a bank not covered by the LCR submits a tender if the operation includes the EWF. In summary, the results in this table{{p}}suggest that banks respond to the LCR by increasing their participation in TDF operations.{{p}}References{{p}}Lee, David S., and Thomas Lemieux, (2010). "Regression Discontinuity Designs in Economics." Journal of Economic Literature, vol. 48{{p}}(2), pp. 281-355.{{p}}1. We thank James Clouse, Francisco Covas, Jeff Huther, and Beth Klee for comments. We also thank Johnathan Hamburg and Cindy Vojtech for kindly{{p}}providing the data on the Term Deposit Facility and on high-quality liquid assets, respectively, and Becky Zhang for excellent research assistance. Return to{{p}}text{{p}}2. An institution can participate in TDF operations if it is eligible to receive interest from the Federal Reserve and if it has a location to settle funds from TDF{{p}}transactions. Interest eligibility is defined by Regulation D. Institutions eligible to receive interest from the Federal Reserve include commercial banks, thrifts,{{p}}and credit unions. Return to text{{p}}3. Depository institutions hold two types of balances in their reserve accounts at the Federal Reserve: balances held to satisfy federally mandated reserve{{p}}balance requirements (RBR) and excess balances or balances above those held to satisfy RBR. Return to text{{p}}4. Level 1 is the highest category of HQLA in the U.S. LCR rule. Level 1 HQLA are not subject to a haircut or to a cap in the LCR calculation. Level 2A and 2B{{p}}HQLA, the other two categories, are subject to haircuts and caps. For more information, please refer to Board of Governors of the Federal Reserve System,{{p}}Federal Deposit Insurance Corporation, Office of the Comptroller of the Currency (2014), "Liquidity Coverage Ratio: Liquidity Risk Management Standards,"{{p}}final rulemaking (Docket No. R-1466), Federal Register, vol. 79 (October 10), pp. 61445-61541. Return to text{{p}}5. Operation details are available on the Federal Reserve Board's website under Monetary Policy and Policy Tools. Return to text{{p}}6. Our main results also hold when we include all domestic depository institutions in the data. In this case, however, we cannot include HQLA data in our{{p}} FRB: FEDS Notes: The Effects of Liquidity Regulation on Participation i...{{p}}4 of 5 6/22/2015 1:36 PM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}regression analysis. These results are available from the authors upon request. Return to text{{p}}7. The term "reports of condition and income" refers to the Consolidated Report of Condition and Income (FFIEC 031 and FFIEC 041). HQLA is not directly{{p}}reported in those forms. The details of how to calculate HQLA are available upon request. Return to text{{p}}8. In addition, LCR banks submit tenders relatively larger than non-LCR banks in operations with an EWF. The average tender amount submitted by a{{p}}non-LCR bank in operations with an EWF, equal to $2 million, is twice as large as the average tender submitted in operations without an EWF, equal to $1{{p}}million; meanwhile, the average tender submitted by an LCR bank in operations with an EWF, equal to $3 billion, is more than three times larger than the{{p}}average tender submitted in operations without an EWF, equal to $978 million. However, in this note we restrict our analysis to participation in TDF operations,{{p}}as opposed to tender amounts, because tender amounts are more sensitive to the effects of the maximum award amounts allowed in TDF operations. Return{{p}}to text{{p}}9. We divide pairs into intervals equal to 0.2ln(7) to ensure that each interval contains observations from at least four different banks. This reduces the effects{{p}}of noise on the average participation rates shown in figure 1. In addition, it keeps information on participation at the bank level confidential. The size of this{{p}}interval is arbitrary and matters only for presentation of the data in this figure. These intervals are not used in the econometric analysis in this memo. Return to{{p}}text{{p}}10. To estimate these polynomials we restrict the sample to observations from banks with assets between $2 billion and $1,250 billion. However, the chart{{p}}does not show observations from banks with assets above $350 billion to maintain the information on participation confidential. In fact, given that there are only{{p}}five banks with assets above $350 billion, points in this chart above this value could reveal participation rates at the bank level. Return to text{{p}}11. Notice that φj and f ( ln( Aij) − ln(50 bi) ) would absorb the effects of EWFj and I( Aij ≥ 50bi ), respectively, in equation (1), thereby making EWFj and I( Aij ≥{{p}}50bi ) redundant in this equation. Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: June 19, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: The Effects of Liquidity Regulation on Participation i...{{p}}5 of 5 6/22/2015 1:36 PM
    Date: 2015–06–19
  23. By: Cukierman, Alex
    Abstract: Persistent decreases in interest rates since the beginning of the twenty first century and the intensification of this trend with the onset of the global financial crisis nurtured the view that the natural rate is substantially lower than it used to be, and by some estimates, even persistently negative. Although investment activity depends mainly on risky rates existing estimates of the natural rate focus mainly on estimation of natural (mostly short term) riskless rates. Gilchrist and Zakrajsek (2012) find that, particularly during crisis times, risky and riskless rates tend to move in opposite directions and that the spread between risky and riskless rates is a good predictor of subsequent economic activity. Drawing on those findings the paper makes a case for conceptualizing and estimating a risky natural rate. This rate which better reflects the impact of the financial system on economic activity, is practically always bounded away from the zero lower bound. After documenting and reviewing the downward trend in world interest rates and the reasons underlying it the paper argues that recent post crisis estimates of the riskless natural rate are likely to be biased downward. Recent estimate of the (unobservable) natural rate are obtained by applying either the Kalman filter or Bayesian estimation to alternative standard versions of the New Keynesian (NK) model. The crisis substantially increased the tightening impact of credit rationing on the New Keynesian (NK) IS relation and the relative importance of the financial stability motive in the monetary policy rule. Since the standard NK model abstracts from credit rationing and from the financial stability motive existing estimates of the natural rate are likely to be biased downward, particularly so since the onset of the crisis.
    Keywords: downward bias in natural rate estimates; risky natural rate
    JEL: E3 E4 E5 G1
    Date: 2016–08
  24. By: Nikola Mirkov; Igor Pozdeev; Paul Söderlind
    Abstract: We ask whether the markets expected the Swiss National Bank (SNB) to discontinue the 1.20 cap on the Swiss franc against the euro in January 2015. In the runup to the SNB announcement, neither options on the euro/Swiss franc nor FX liquidity indicated a significant shift in market expectations. Furthermore, we find that the SNB's verbal interventions during the period of cap enforcement increased the credibility of the cap by reducing the uncertainty of future euro/Swiss franc rate. We conclude that the markets did not anticipate the discontinuation of the policy.
    Keywords: Swiss franc, implied volatilities, market expectations
    JEL: E58 E44 G12
    Date: 2016
  25. By: Ellen E. Meade; Nicholas A. Burk; Melanie Josselyn
    Abstract: May 26, 2015{{p}}The FOMC meeting minutes: An assessment of counting words and the diversity of views{{p}}Ellen E. Meade, Nicholas A. Burk, and Melanie Josselyn1{{p}}The Federal Reserve's communications with the public have evolved substantially since the early 1990s and today include: policy{{p}}statements released shortly after the conclusion of monetary policy meetings; minutes of those meetings issued three weeks later;{{p}}quarterly economic forecasts from the members of the Federal Reserve Board of Governors and the presidents of the Federal Reserve{{p}}Banks; the Chair's press conferences four times per year; a semi-annual Monetary Policy Report that is submitted to the Congress and{{p}}released to the public, along with the Chair's testimony on that report; and transcripts of monetary policy meetings published after five{{p}}years. In addition, the public websites maintained by the Board and the 12 Federal Reserve Banks provide reports, testimony, speeches,{{p}}and a wealth of other information on the policy and operations of the Federal Reserve System.{{p}}In this note, we focus on the minutes of Federal Open Market Committee (FOMC) meetings. Summaries of FOMC meetings have been{{p}}released to the public in some form since 1936; detailed minutes like those available today have been released since 1993. Between{{p}}1993 and the end of 2004, minutes for each meeting were released after the subsequent meeting and so did not inform the public about{{p}}the most recent thinking or discussions of the FOMC. In December 2004, the Committee decided to begin publishing the minutes 21{{p}}days after each FOMC meeting, shortening the publication lag by several weeks and providing the public with more timely information{{p}}about policy deliberations and the rationale for policy decisions.{{p}}The minutes provide a detailed summary of the discussion at an FOMC meeting.2 Typically, an FOMC meeting begins with a staff review{{p}}of foreign currency and domestic open market operations over the intermeeting period; meeting participants have the opportunity to{{p}}question the staff on these market developments. Next comes the "economic go-round," which begins with several staff presentations on{{p}}developments in, and prospects for, the economies of the United States and foreign countries. After asking any questions they may have{{p}}on those presentations, all meeting participants--the Federal Reserve governors and the presidents of the 12 Federal Reserve Banks--{{p}}discuss their views on economic developments and the outlook. Generally, Reserve Bank presidents include in their remarks{{p}}commentary from industry contacts and information about recent economic developments in their Districts. In addition, four times each{{p}}year, the economic go-round includes a staff summary of the economic projections submitted by FOMC participants.3{{p}}A go-round on monetary policy follows the economic go-round. A staff briefing outlines policy options and discusses several alternatives{{p}}for the statement that the Committee will issue after the meeting. Following questions, the governors and Reserve Bank presidents{{p}}outline their views on policy and discuss possible amendments to the language in drafts of the Committee's statement. At the conclusion{{p}}of that discussion, the Committee members vote on the policy that will be followed over the period until the next FOMC meeting and on{{p}}the statement that will be released shortly after the meeting is adjourned. A key feature of the governance of Federal Reserve monetary{{p}}policy decisions is that, while all Reserve Bank presidents participate in the discussion at FOMC meetings, not all are voting members of{{p}}the Committee. The Federal Reserve Act specifies that Committee "members"--that is, those who vote on policy--include all Federal{{p}}Reserve governors, the President of the Federal Reserve Bank of New York, and four of the other 11 Federal Reserve Bank presidents{{p}}who vote according to a specified rotation.{{p}}In addition to the review of financial market developments and the go-rounds on the economy and monetary policy, many FOMC{{p}}meetings include discussion of a special topic. In recent years, for example, meeting participants have discussed the Federal Reserve's{{p}}large-scale asset purchase programs, the forward guidance about future policy that is contained in the Committee's postmeeting{{p}}statement, and the consensus statement on the Committee's longer-run goals and monetary policy strategy that was first issued in{{p}}January 2012. The minutes of FOMC meetings summarize participants' discussions of such special topics as well as their comments{{p}}during the regular go-rounds.{{p}}The role of the minutes and the postmeeting statement{{p}}Transcripts from FOMC meetings around the time that the Committee decided to expedite the release of the minutes shed some light on{{p}}the role that policymakers see the minutes playing in their communications about monetary policy, particularly relative to the postmeeting{{p}}statement that the Committee releases to the public shortly after the conclusion of each FOMC meeting. In May 2005, Reserve Bank{{p}}Presidents Guynn (Atlanta) and Poole (St. Louis) noted their increasing discomfort with the fact that the statement did not reflect the{{p}}widespread uncertainty that meeting participants were expressing about policy at that meeting. They wanted that range of uncertainty to{{p}}be presented in the statement, "because, otherwise, the statement would not truly reflect what happened at the meeting."4 Chairman{{p}}Greenspan responded that "what is in the minutes, even though they are released later, reflects what occurred before we voted on the{{p}}statement. So whatever the Committee votes is the Committee's view at that point."5 At a meeting in 2006, Chairman Bernanke framed{{p}}the relative roles of the statement and the minutes slightly differently, saying, "... I do not think that the minutes and the statement are{{p}}perfect substitutes. The statement, after all, is much more timely, and it represents something closer to a consensus or median view of{{p}} FRB: FEDS Notes: The FOMC meeting minutes: An assessment of coun...{{p}}1 of 5 5/26/2015 1:37 PM{{p}}Table 1. Quantitative words used in the FOMC{{p}}minutes{{p}}"all"{{p}}"all but one"{{p}}"almost all"{{p}}"most"{{p}}"many"{{p}}"several"{{p}}"few"{{p}}"a couple" or "two"{{p}}"one" ... "another{{p}} Source: Deborah J. Danker and Matthew M. Luecke (2005),{{p}}"Background on FOMC Meeting Minutes," Federal Reserve{{p}}Bulletin, Spring, pp. 175-179.{{p}}the Committee as opposed to the minutes, which try to express the range of views and discussion around the table."6 These extracts{{p}}from FOMC transcripts suggest that policymakers see the minutes as providing insight about the breadth of views that the postmeeting{{p}}statement does not provide.{{p}}All of the minutes released since the end of 2004 on the expedited schedule have reflected not only views expressed by Committee{{p}}"members" but also those articulated by meeting "participants," a larger group that includes those Reserve Bank presidents who do not{{p}}vote. The minutes' account of the economic go-round summarizes the views expressed by participants, while the section titled{{p}}"Committee Policy Action" includes only the members because they are responsible for the policy decision. At times since 2008, the{{p}}minutes have included participants' views on particular aspects of monetary policy when the range of views expressed by participants{{p}}was broader than that expressed by the members alone.7{{p}}Another way in which the meeting minutes are informative about the diversity of views is through the use of "quantitative" or "counting"{{p}}words--such as "few" or "many"--to characterize the number of members or participants aligned with a particular view. In their article on{{p}}the minutes, Danker and Luecke (2005) provide a list of the quantitative words, which are shown on Table 1 ordered from largest to{{p}}smallest. We add to this list two other counting words that are used with some frequency in the minutes: "some" and "a number of."{{p}}An analysis of the "counting" words{{p}}The question we are interested in is, do the FOMC minutes present diverse{{p}}perspectives and has that diversity changed over time? To find the answer, we{{p}}subject the FOMC minutes released from 2005 through 2014 to some{{p}}elementary text analysis and provide an assessment of the dispersion of{{p}}viewpoints expressed through an examination of the quantitative words. The{{p}}frequency of these counting words can be used as a measure of the range of{{p}}opinion expressed at an FOMC meeting.{{p}}For the purposes of this analysis, we examined all sections of the minutes that{{p}}describe policymakers' discussion at the meeting; these are, primarily, the{{p}}sections on the economic and monetary policy go-rounds titled "Participants'{{p}}Views on Current Conditions and the Economic Outlook" and "Committee Policy{{p}}Action," respectively, although we also examined any special sections or{{p}}additional paragraphs that included sentences pertaining to views expressed by{{p}}participants.8{{p}}Before turning to our findings, we note a few details about our counting procedures. First, statements that represent unanimity are{{p}}sometimes expressed explicitly, as in "all members judged;" at other times, statements of near unanimity are implicit, as in "participants{{p}}judged" or "the Committee judged." In making the counts, we grouped these explicit and implicit statements together with statements{{p}}using "all but one" and "almost all" to form a grouping that reflects a high degree of consensus.9 Second, the quantitative word "one" is{{p}}sometimes used in a "one ... another" construction; when that occurred, we counted both "one" and "another" as separate views. Finally,{{p}}we counted only expressions of views or opinions, and not mere discussions of a topic.{{p}}It is important to note that the FOMC minutes have gotten considerably longer over time. In 2005, the first year that the minutes were{{p}}released on their current, expedited schedule, the average length per meeting was just under 4200 words.10 By 2014, the word count{{p}}had risen to around 8350, with the bulk of this rise occurring between 2005 and 2009. The total annual frequency of the counting words{{p}}from participants' and members' paragraphs of the minutes has risen over time as well, particularly since about 2011 (chart 1).{{p}}Interestingly, the counting words have risen as a share in the total number of words (chart 2), pointing either to a greater diversity of{{p}}viewpoints or to more complete reporting of the diversity of views.{{p}}Chart 1: Total quantitative words, FOMC minutes, 2005-2014{{p}} FRB: FEDS Notes: The FOMC meeting minutes: An assessment of coun...{{p}}2 of 5 5/26/2015 1:37 PM{{p}}Accessible version{{p}}Chart 2: Quantitative words as a share of total words, annual average{{p}}Accessible version{{p}}The total number of paragraphs that reflect views expressed by members is substantially smaller than the total number of paragraphs{{p}}describing participants' views--sometimes substantially so. Thus, it is not surprising that the total count for the quantitative words is{{p}}higher for participants' paragraphs than for members' paragraphs (not shown). Since 2007, the quantitative words in participants'{{p}}paragraphs of the minutes have accounted for about 75 to 80 percent of all quantitative words used.{{p}}Chart 3 provides the share of each counting word in the total for participants (upper panel) and members (lower panel) for each year.{{p}}Highly consensual statements (shown as "consensus") comprised nearly 60 percent of participants' paragraphs in 2005, compared with{{p}}about 25 percent in 2014. Members' paragraphs have a higher proportion of "consensus" statements, which is to be expected in light of{{p}}the role that the Committee's statement and the discussion of it plays in the members' paragraphs. In 2014, "consensus" statements{{p}}constituted 80 percent of members' paragraphs, although for several years prior the share was somewhat lower--generally between 60{{p}}and 70 percent. A wider diversity of opinion among participants and members might be expected in the aftermath of the financial crisis{{p}}given that, after the Committee reduced the target for the federal funds rates to nearly zero in December 2008, the Federal Reserve{{p}}began using nontraditional monetary policy--namely, large-scale asset purchases and forward guidance--to provide additional stimulus{{p}} FRB: FEDS Notes: The FOMC meeting minutes: An assessment of coun...{{p}}3 of 5 5/26/2015 1:37 PM{{p}}to the economy. These nontraditional monetary policies were less familiar and were therefore less well understood. Policymakers{{p}}debated their efficacy and costs and spent a great deal of time at FOMC meetings discussing the implications and effects of them. In{{p}}addition, the greater diversity of viewpoints could reflect a changing composition of policymakers with different opinions.11{{p}}Chart 3: Shares of counting words in total, FOMC participants (upper) and members (lower){{p}}Accessible version{{p}}In conclusion, the FOMC meeting minutes are a key means for informing the Congress and the public about the full range of{{p}}policymaker opinion and debate about monetary policy issues, and thus help ensure that the Federal Reserve is accountable to the{{p}}Congress and the public. Since 2005, the minutes have gotten considerably longer. Our analysis of the quantitative words used in the{{p}}minutes indicates that the meeting minutes capture a wide diversity of viewpoints expressed by FOMC policymakers and that this{{p}}diversity appears to have increased over time, particularly since the financial crisis.{{p}}1. The authors thank Bill English, Thomas Laubach, Steve Meyer, and Bob Tetlow for comments. Return to text{{p}}2. For a more detailed discussion of what happens at an FOMC meeting, see the speech "Come with Me to the FOMC" that Governor Elizabeth A. Duke gave{{p}}to the Money Marketeers of New York University in October 2010. Return to text{{p}}3. These projections have been collected since October 2007. Tables and charts of the projections are released to the public prior to the Chair's press{{p}} FRB: FEDS Notes: The FOMC meeting minutes: An assessment of coun...{{p}}4 of 5 5/26/2015 1:37 PM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}conference and a detailed analysis is released three weeks later in the "Summary of Economic Projections" section of the minutes for the relevant FOMC{{p}}meetings. Return to text{{p}}4. President Poole, Transcript of FOMC meeting, May 3, 2005, p. 86. Return to text{{p}}5. Chairman Greenspan, Transcript of FOMC meeting, May 3, 2005, p. 86. Return to text{{p}}6. Chairman Bernanke, Transcript of FOMC meeting, March 27-28, 2006, p. 140. Return to text{{p}}7. For example, in the minutes of the January 2013 meeting, there were three paragraphs at the end of the economic go-round section on the benefits and{{p}}costs of the Committee's open-ended asset purchase program, the pace of asset purchases, and the economic thresholds in the Committee's forward{{p}}guidance. On some occasions, there have been separate sections in the minutes pertaining to a discussion of aspects of monetary policy; see, for example,{{p}}the section titled "Policy Planning" in the minutes of the October 2013 meeting which relates participants' views on the strategy and tactics of future policy.{{p}}Return to text{{p}}8. The "Committee Policy Action" section reflects views expressed by members; it does not reflect views expressed only by those participants who are not{{p}}voting members. Return to text{{p}}9. We also included statements using the construct "participants generally judged..." or "members generally judged..." in this grouping. Return to text{{p}}10. The word counts and other analysis in this section include scheduled meetings and exclude video conference (unscheduled) meetings. The minutes of the{{p}}first meeting in a year are significantly longer than subsequent meetings owing to discussion of procedural matters. Return to text{{p}}11. For example, since the beginning of 2005, there has been considerable turnover among Federal Reserve governors: Six governors appointed prior to 2005{{p}}resigned and, of the 12 governors appointed since 2005, only five of them remain.Return to text{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: May 26, 2015{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: The FOMC meeting minutes: An assessment of coun...{{p}}5 of 5 5/26/2015 1:37 PM
    Date: 2015–05–26
  26. By: Boubaker, Sabri; Gounopoulos, Dimitrios; Nguyen, Duc Khuong; Paltalidis, Nikos
    Abstract: US public pension funds deficits remain stubbornly high even though market conditions have improved in the post-crisis period. This article examines the role of lower short- and long-term interest rates imposed by the use of unconventional monetary policy on pension funds risk taking and asset allocation behavior. We quantify the effects of the Zero Lower Bound policy and the launch of unconventional monetary policy measures by using two structural Vector AutoRegression (VAR) models, a Bayesian VAR and a Markov switching-structural VAR. We provide the first comprehensive evidence showing that persistently low interest rates and falling Treasury yields cause a substantial increase in pension funds risk and portfolios beta. Additionally, we document that the severe funding shortfall in many pension schemes is, to a large extent, associated with and prompted by changes in the monetary policy framework.
    Keywords: Pension funds; Unconventional monetary policy; Asset allocation; Zero lower bound
    JEL: E52 G11 G23
    Date: 2015–05
  27. By: Timothy S. Hills; Taisuke Nakata; Sebastian Schmidt
    Abstract: Print{{p}}February 12, 2016{{p}}The Risk of Returning to the Effective Lower Bound: An Implication for Inflation Dynamics{{p}}after Lift-Off 1{{p}}Timothy Hills, Taisuke Nakata, and Sebastian Schmidt2{{p}}1. Introduction{{p}}In this note, we analyze an implication of the effective lower bound (ELB) risk—the possibility that adverse shocks will force{{p}}policymakers in the future to lower the policy rate to the ELB—on inflation dynamics after liftoff. The implication we analyze is{{p}}deflationary bias—a phenomenon in which the tail risk induced by the ELB constraint leads inflation to decline under conventional{{p}}monetary policy rules. In particular, our focus is examining how large the deflationary bias is at the economy's risky steady state—the{{p}}point to which the economy eventually converges as headwinds and tailwinds dissipate.{{p}}We first document, briefly, how academic economists, policymakers, and financial market participants assess the ELB risk. We then use{{p}}an empirically rich DSGE (Dynamic Stochastic General Equilibrium) model to analyze how large the deflationary bias induced by the{{p}}ELB risk might be in the U.S. Our baseline simulation suggests that the ELB risk causes inflation to undershoot the target rate of 2{{p}}percent by about 20 basis points at the economy's risky steady state under an empirically plausible specification of monetary policy. We{{p}}find that the deflationary bias induced by the ELB risk can increase to as large as 40 basis points under alternative plausible{{p}}assumptions regarding the long-run equilibrium real rate.{{p}}Overall, our result provides a cautionary tale for policymakers aiming to raise inflation from currently low levels: Even after liftoff, the ELB{{p}}constraint can have enduring adverse effects on inflation through its effects on expectations. The note concludes by briefly discussing{{p}}some implications of this enduring effect of the ELB constraint for monetary policy strategy. The analyses in this note draw heavily on{{p}}Hills, Nakata, and Schmidt (2016).{{p}}2. ELB risk after lift-off{{p}}Academic economists and policymakers have noted that the ELB is not likely to be a one-time event. Ball (2013) argues that, if the{{p}}inflation target remains 2 percent, "the lower bound on interest rates is likely to constrain monetary policy in a large fraction of{{p}}recessions" in the U.S. In its April 2014 WEO, the IMF extensively studies the world wide decline in equilibrium real rates and states that{{p}}the probability of hitting the ELB is likely to increase going forward (IMF (2014)). A renewed interest in the analysis of optimal inflation{{p}}targets and alternative monetary policy frameworks is based on the assumption that the ELB can bind again (Blanchard et al. (2010) and{{p}}Bernanke (2015)).{{p}}Financial market participants also see the ELB binding again in the future as a quantitatively significant possibility. According to a special{{p}}question in the Survey of Primary Dealers on December 2015, a median respondent attached 20 percent probability to the event that the{{p}}federal funds rate returns to the ELB within two years after lift-off.{{p}}3. An implication of the ELB risk: Deflationary bias{{p}}3.1. What is deflationary bias?{{p}}When firms are forward-looking in their pricing decisions, the ELB risk leads inflation to fall below the inflation target at the economy's{{p}}risky steady state, provided that monetary policy is characterized by a standard interest-rate feedback rule. In plain English, this means{{p}}that inflation will return to a level below the inflation target when all headwinds or tailwinds dissipate. This also means that inflation{{p}}fluctuates around that level even in absence of actual ELB episodes. This phenomenon occurs because forward-looking firms base their{{p}}pricing decisions on the expected economic conditions in the future, and the tail risk induced by the ELB constraint lowers the expected{{p}}economic conditions. In the academic literature, this phenomenon is referred to as deflationary bias.3{{p}}3.2. How large is the deflationary bias?{{p}}We use an empirically rich DSGE model calibrated to broadly match key moments of the output gap, inflation, and the federal funds rate{{p}}in the U.S. over the last two decades to get some sense of how large the deflationary bias may be in reality. The model is a standard{{p}}New Keynesian model augmented with consumption habits in the household's preference, sticky wages, and an interest-rate smoothing{{p}}term in the policy rule. There are two shocks---productivity and demand shocks. The main shock that drives fluctuations in this economy{{p}}is the demand shock, which is implemented through time-varying discount rates.4{{p}}The inflation target in the model's interest-rate feedback rule is set to 2 percent. The steady-state discount rate and the growth rate of{{p}}total factor productivity are chosen so that the deterministic steady-state policy rate is 3.75 percent. The effective lower bound on the{{p}}federal funds rate is set to 13 basis points. The list of parameter values is in Hills, Nakata, and Schmidt (2016). The model is solved{{p}}globally using a nonlinear solution method.5{{p}} FRB: FEDS Notes: The Risk of Returning to the Effective Lower Bound...{{p}}1 of 6 2/12/2016 1:20 PM{{p}}The first three rows in Table 1 show the unconditional standard deviations of inflation, the output gap and the nominal interest rate from{{p}}the model and from the data. The standard deviations of the output gap, inflation, and the policy rate in the model are very close to those{{p}}in the data (3.1, 0.42, and 2.34 in the model versus 2.9, 0.52, and 2.34 in the data). The next two rows in Table 1 show the conditional{{p}}mean of the output gap and inflation when the policy rate is at the ELB from the model and from the data. The conditional means of the{{p}}output gap and inflation are somewhat higher and lower than those in the data, but are reasonably close. Finally, the frequency of being{{p}}at the ELB and the expected duration of the ELB episode are 16 percent and 9 quarters in the model, versus 36 percent and 26 quarters{{p}}in the data.6{{p}}Table 1: Key Moments{{p}}Moments Variable Model Data (1995Q3-2015Q2){{p}}St.Dev (X) Output Gap 3.1 2.9{{p}}Inflation 0.42 0.52{{p}}Policy Rate 2.34 2.34{{p}}E(X|ELB) Output Gap -3.4 -4.2{{p}}Inflation 1.18 1.48{{p}}Policy Rate 0.13 0.13{{p}}ELB Frequency 16% 36%{{p}}Expected/Actual Duration 9 quarters 26 quarters{{p}}Table 2 compares the deterministic and risky steady states of inflation, the output gap, and the policy rate. Due to the deflationary bias{{p}}explained earlier, inflation is lower at the risky steady state than at the deterministic steady state Inflation is about 30 basis points lower{{p}}at the risky steady state than the target rate of 2 percent. The policy rate is lower at the risky steady state than the deterministic steady{{p}}state (3.04 percent versus 3.75 percent), as a lower inflation is associated with a lower policy rate in the interest-rate feedback rule.7 The{{p}}expected real rate is lower at the risky steady state, and thus the output gap is positive at the risky steady state, albeit slightly.{{p}}Table 2: The Effects of the ELB Risk on the Steady State{{p}}Inflation Output Gap Policy Rate{{p}}Deterministic Steady State 2 0 3.75{{p}}Risky Steady State 1.71 0.32 3.04{{p}}(wedge) (-0.29) (0.32) (-0.71){{p}}E(X) 1.66 0.29 2.8{{p}}Risky Steady State (No ELB Constraint) 1.88 0.04 3.37{{p}}(wedge) (-0.12) (0.04) (-0.38){{p}}Since the ELB constraint is not the only nonlinear feature of the model, there is some difference between the deterministic and risky{{p}}steady states even in the absence of the ELB constraint, as shown in the last two rows of Table 2. For inflation, the risky steady state is{{p}}1.88 percent in the model without the ELB constraint. The contribution of the ELB risk to the overall deflationary bias at the risky steady{{p}}state is 17 basis points. Thus, the majority of the overall deflationary bias in the model with the ELB constraint comes from the ELB{{p}}constraint, instead of other nonlinearities of the model.{{p}}Note that the risky steady state is conceptually different from the average. Let's take inflation as an example. The risky steady state{{p}}inflation is the point around which inflation fluctuates, while the average inflation is the average of inflation in all states of the economy.{{p}}As shown in Table 2 and Figure 1, the risky steady state inflation is higher than the average inflation in the economy with the ELB{{p}}constraint, as the ELB constraint makes the unconditional distribution of inflation negatively skewed. The observation that the ELB{{p}}constraint pushes down the average inflation below the target by creating asymmetry in the distribution of inflation is intuitive and has{{p}}been well known for a long time (Coenen, Orphanides, and Wieland (2004) and Reifschneider and Williams (2000)). This holds true{{p}}even when price-setters form expectations in a backward-looking manner. The result that the ELB risk lowers the center of the{{p}}distribution below the target is less intuitive and requires that price-setters are forward-looking in forming their expectations.{{p}}Figure 1: Unconditional Distribution of Inflation{{p}} FRB: FEDS Notes: The Risk of Returning to the Effective Lower Bound...{{p}}2 of 6 2/12/2016 1:20 PM{{p}} Note: Authors' calculation as described in Hills, Nakata, and Schmidt (2016){{p}}Accessible version{{p}}To emphasize the fact that the possibility of the ELB is the source of the deflationary bias, we contrast the impulse response functions of{{p}}inflation, output, and the policy rate from the model with those from a perfect-foresight version of the model, which ignores the ELB risk.{{p}}Figure 2 shows the impulse response functions when the initial demand and productivity shocks are chosen so that inflation is 50 basis{{p}}points, the output gap is -7 percent, and the policy rate is at the ELB at period one. Black lines are from the perfect-foresight version of{{p}}the model that abstracts from the ELB risk, while red lines are from the model that correctly takes the ELB risk into account. Under{{p}}perfect-foresight, the policy rate stays at the ELB for 11 quarters and will gradually converge to 3.75 percent. Inflation slowly rises to 2{{p}}percent, and the output gap will converge to zero after some overshooting. On the other hand, when the private sector correctly{{p}}acknowledges the ELB risk, the policy rate stays at the ELB for 13 quarters and returns to its risky steady state of about 3 percent.{{p}}Inflation slowly rises to about 1.7 percent and the output gap remains slightly positive in the long run.{{p}}Figure 2: The Effect of the ELB Risk on Projections{{p}} FRB: FEDS Notes: The Risk of Returning to the Effective Lower Bound...{{p}}3 of 6 2/12/2016 1:20 PM{{p}} Note: Authors' calculation as described in Hills, Nakata, and Schmidt (2016){{p}}Accessible version{{p}}3.3. The long-run equilibrium interest rate and deflationary bias{{p}}The magnitude of the deflationary bias depends importantly on the frequency of the policy rate being at the ELB, which in turn depends{{p}}on the long-run equilibrium policy rate. Figure 3 shows how the magnitude of deflationary bias depends on the (deterministic){{p}}steady-state level of the policy rate. In this figure, we vary the parameter governing the long-run growth rate of labor productivity to{{p}}induce the change in the steady-state policy rate. As demonstrated in the top-right panel, deflationary bias is larger when the{{p}}deterministic steady state policy rate is lower. This is because the probability of hitting the ELB is higher with a lower steady-state policy{{p}}rate, as shown in the top-left panel of the figure. When the deterministic steady state policy rate is 3.35 percent, the probability of being{{p}}at the ELB is 28 percent and inflation is about 50 basis points below the inflation target at the economy's risky steady state. Comparing{{p}}this deflationary bias with that in the model without the ELB constraint, indicated by the dashed black line, shows that the ELB risk{{p}}accounts for about 40 basis points of the overall deflationary bias.8{{p}}Figure 3: Long-Run Policy Rate and Deflationary Bias{{p}} FRB: FEDS Notes: The Risk of Returning to the Effective Lower Bound...{{p}}4 of 6 2/12/2016 1:20 PM{{p}} Note: Authors' calculation as described in Hills, Nakata, and Schmidt (2016){{p}}Accessible version{{p}}4. Concluding remarks{{p}}This note has demonstrated that the deflationary bias induced by the ELB risk is non-trivial in an empirically rich DSGE model calibrated{{p}}to match key features of the U.S. economy. The result that the ELB constraint has enduring effects on the economy even after liftoff has{{p}}important implications for the design of monetary policy. For example, the deflationary bias, while typically ignored in the analyses of{{p}}optimal inflation targets, importantly affects the cost-benefit calculation of changing the inflation target. As another example, Nakata and{{p}}Schmidt (2014) show that the deflationary bias makes it desirable for the central bank to put more weight on the inflation stabilization{{p}}objective, relative to the output stabilization objective.9{{p}}This conclusion is of course conditional on the validity of the model used for the analysis. While the model is a variant of the standard{{p}}model widely used at central banks, there is a continuing debate about the usefulness of this model for policy analysis among{{p}}economists. Thus, caution is needed when drawing policy implications.10{{p}} References:{{p}}Adam, K., and R. Billi (2007): "Discretionary Monetary Policy and the Zero Lower Bound on Nominal Interest Rates," Journal of{{p}}Monetary Economics, 54(3), 728-752.{{p}}Ball, L. M. (2013): "The Case for Four Percent Inflation," Central Bank Review, 13, 17-31.{{p}}Bernanke, B. (2015): "Monetary Policy in the Future," Remarks at the Rethinking Macro Policy III Conference at the IMF, Washington,{{p}}DC.{{p}}Blanchard, O., G. Dell'Ariccia, and P. Mauro (2010): "Rethinking Macroeconomic Policy," Journal of Money, Credit, and Banking, 42(s1),{{p}}199-215.{{p}}Chung, H., E. Herbst, and M. T. Kiley (2014): "Effective Monetary Policy Strategies in New Keynesian Models: A Re-Examination,"{{p}}NBER Working Papers 20611, National Bureau of Economic Research.{{p}}Coenen, G., A. Orphanides, and V. Wieland (2004): "Price Stability and Monetary Policy Effectiveness when Nominal Interest Rates are{{p}}Bounded at Zero," B.E. Journal of Macroeconomics: Advances in Macroeconomics, 4(1).{{p}} FRB: FEDS Notes: The Risk of Returning to the Effective Lower Bound...{{p}}5 of 6 2/12/2016 1:20 PM{{p}}Accessibility Contact Us Disclaimer Website Policies FOIA PDF Reader{{p}}Evans, C., J. Fisher, F. Gourio, and S. Krane (2015): "Risk Management for Monetary Policy Near the Zero Lower Bound," Brookings{{p}}Papers on Economic Activity Conference Draft.{{p}}Hills, T. S., T. Nakata, and S. Schmidt (2016): "The Risky Steady State and the Interest Rate Lower Bound," Finance and Economics{{p}}Discussion Series 2016-009, Board of Governors of the Federal Reserve System (U.S.).{{p}}IMF (2014): "World Economic Outlook" April 2014{{p}}Nakata, T., and S. Schmidt (2014): "Conservatism and Liquidity Traps," Finance and Economics Discussion Series 2014-105, Board of{{p}}Governors of the Federal Reserve System (U.S.).{{p}}Nakov, A. (2008): "Optimal and Simple Monetary Policy Rules with Zero Floor on the Nominal Interest Rate," International Journal of{{p}}Central Banking, 4(2), 73-127.{{p}}Reifschneider, D., and J. C. Williams (2000): "Three Lessons for Monetary Policy in a Low-Inflation Era," Journal of Money, Credit and{{p}}Banking, 32(4), 936-966.{{p}}1. We would like to thank Jean-Phillipe Laforte and John Roberts for their comments. Paul Yoo provided excellent research assistance. The views expressed in{{p}}this note, and all errors and omissions, should be regarded as those solely of the author, and are not necessarily those of the Federal Reserve Board of{{p}}Governors, the Federal Reserve System, or the European Central Bank. Return to text{{p}}2. Timothy Hills is a Ph.D. student at the Stern School of Business, New York University; Taisuke Nakata is an economist at the Board of Governors of the{{p}}Federal Reserve System (Division of Research and Statistics); Sebastian Schmidt is an economist at the European Central Bank (Monetary Policy Research{{p}}Division). Return to text{{p}}3. The implications of deflationary bias for optimal policy have been studied in a stylized New Keynesian model by Adam and Billi (2007), Evans et al. (2015),{{p}}Nakata and Schmidt (2014), and Nakov (2008). Note that the magnitude of the deflationary bias depends on the policy rule in place. Under the price-level{{p}}targeting or optimal commitment policy, the deflationary bias would be smaller as the decline in inflation at the ELB would be smaller. Return to text{{p}}4. See Hills, Nakata, and Schmidt (2016) for more details on the model. Return to text{{p}}5. Due to the computational burden of globally solving dynamic models, economists typically solve monetary DSGE models under the assumption of perfect-foresight.{{p}}That is, after lift-off, the households and firms are assumed to attach zero probability to the event that adverse shocks will push the policy rate back{{p}}to the ELB. Due to curse of dimensionality, the computational burden of globally solving DSGE models increases exponentially with the size of the model. It is{{p}}nearly infeasible to solve more empirically-rich structural models--such as FRB/US, EDO, and SIGMA used at the Federal Reserve--in a reasonable amount of{{p}}time without perfect-foresight assumptions. Return to text{{p}}6. The ELB frequency in the data depends importantly on the sample period. We use the last two decades as our sample because the long-run inflation{{p}}expectations were low and stable during this period. Return to text{{p}}7. As we discuss in Hills, Nakata, and Schmidt (2016), another way to understand why inflation and the policy rate are lower at the risky steady state than at{{p}}the deterministic steady state is to examine the interaction of the Taylor rule with a version of the Fisher relationship modified to account for the effects of risk.{{p}}Return to text{{p}}8. A similar picture emerges if one considers a change in the target rate of inflation or the value of the effective lower bound, as shown in Hills, Nakata, and{{p}}Schmidt (2016). Return to text{{p}}9. In Hills, Nakata, and Schmidt (2016), we study how each parameter in the interest-rate feedback rule affects the magnitude of deflationary bias and show{{p}}that a higher inertia, a higher inflation target, and a lower effective lower bound can reduce the size of the deflationary bias. Return to text{{p}}10. See, for example, Chung, Herbst, and Kiley (2014). Return to text{{p}}Please cite this note as:{{p}}Hills, Timothy, Taisuke Nakata, and Sebastian Schmidt (2016). "The Risk of Returning to the Effective Lower Bound: An Implication for{{p}}Inflation Dynamics after Lift-Off," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, February 12,{{p}}{{p}} Disclaimer: FEDS Notes are articles in which Board economists offer their own views and present analysis on a range of topics in{{p}}economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers.{{p}}Last update: February 12, 2016{{p}}Home | Economic Research & Data{{p}} FRB: FEDS Notes: The Risk of Returning to the Effective Lower Bound...{{p}}6 of 6 2/12/2016 1:20 PM
    Date: 2016–02–12
  28. By: Fernando Alvarez (University of Chicago); Francesco Lippi (Einaudi Institute (EIEF))
    Abstract: Transitory price changes are prominent in the data but do not fit neatly in standard sticky price models. We present a model where a firm chooses a “price plan†, namely a set of 2 prices each of which can be freely posted at each moment. While price changes between prices within the plan are free, the plan can be changed only subject to a fixed menu cost. This setup generates a persistent “reference†price level and short lived deviations from it, as in many datasets and a decreasing hazard function for price changes. We analytically solve for the optimal policy and for the cumulative impulse response function of output to a monetary shock. We compare the economy with the 2-price plan to a menu-cost economy (i.e. a plan with 1 price) featuring the same number of persistent price changes. We show that, for a small monetary shock, the introduction of a plan with 2 prices yields a cumulative output response that is 1/3 of the one produced by the menu cost economy. The smaller real effect is due to the flexibility delivered by the temporary price changes that are used by firms to respond to the shock.
    Date: 2016
  29. By: Thomas Winberry (University of Chicago); Pablo Ottonello (University of Michigan)
    Abstract: Aggregate investment is central to the conduct of monetary both empirically and theoretically. However, existing studies largely ignore the extensive micro-level heterogeneity in investment behavior across firms. In this paper, we reassess the investment channel of monetary policy in the presence of this heterogeneity by empirically studying the response of micro-level investment to monetary shocks, and using these estimates to build a quantitative heterogeneous firm model for policy analysis. Our empirical analysis combines firm-level Compustat investment data with a high-frequency identification of monetary policy shocks. We will then build a quantitative general equilibrium model incorporating the relevant sources of heterogeneity identified in the data. We will use the model to study two broad issues. First, we will compute the average aggregate effect of monetary policy shocks, and compare the results to the estimated VAR literature. Second, we will study how the distribution of underlying heterogeneity shapes the response to monetary shocks at different points in time. How does the aggregate effect of monetary policy depend on the distribution of productivity, capital, or net worth? Does it vary substantially over the cycle? What does all this imply for the design of monetary policy going forward?
    Date: 2016
  30. By: Shuping Shi (Macquarie University); Stan Hurn (QUT); Peter C B Phillips (Yale University)
    Abstract: This paper re-examines changes in the causal link between money and income in the United States for over the past half century (1959 - 2014). Three methods for the data-driven discovery of change points in causal relationships are proposed, all of which can be implemented without prior detrending of the data. These methods are a forward recursive algorithm, a recursive rolling algorithm and the rolling window algorithm all of which utilize subsample tests of Granger causality within a lag-augmented vector autoregressive framework. The limit distributions for these subsample Wald tests are provided. The results from a suite of simulation experiments suggest that the rolling window algorithm provides the most reliable results, followed by the recursive rolling method. The forward expanding window procedure is shown to have worst performance. All three approaches find evidence of money-income causality during the Volcker period in the 1980s. The rolling and recursive rolling algorithms detect two additional causality episodes: the turbulent period of late 1960s and the starting period of the subprime mortgage crisis in 2007.
    Keywords: Time-varying Granger causality, subsample Wald tests, Money-Income
    JEL: C12 C15 C32 E47
    Date: 2016–08–30
  31. By: Corsetti, G.; Mavroeidi, E.; Thwaites, G.; Wolf, M.
    Abstract: We study how small open economies can engineer an escape from deflation and unemployment in a global secular stagnation. Building on the framework of Eggertsson et al. (2016), we show that the transition to full employment requires a dynamic depreciation of the exchange rate, without prejudice for domestic inflation targeting. However, if depreciation has strong income and valuation effects, the escape can be beggar thy self, raising employment but actually lowering welfare. We show that, while a relaxation in the Effective Lower Bound (ELB) can work as a means of raising employment and inflation in financially closed economies, it may have exactly the opposite effect when economies are financially open.
    Keywords: Small open economy, secular stagnation, capital controls, optimal policy, zero lower bound
    JEL: F41 E62
    Date: 2016–08–15

This nep-mon issue is ©2016 by Bernd Hayo. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.
Use to sign off.