Referee Finance

Home of the Financial Intelligence Lab

  • Increase font size
  • Default font size
  • Decrease font size

Bond Yields and the Federal Reserve

E-mail Print PDF

Bond Yields and the Federal Reserve

In her paper “Bond Yields and the Federal Reserve”, Monika Piazzesi introduces a new point of view concerning the relationship between prices of bonds and swaps and decisions about interest rates taken by the Federal Open Market Committee (FOMC). She tries to match the yield curve shape, modelling a two-way process where information available in the market and FOMC decisions influence one another. Previous models did not use an observable target, nor focused on the whole term structure of the yield curve. The new approach that the author uses in this paper is able to explain in a satisfying way the shaped of volatility curve as a combined result of an autoregressive target setting.

The model improves previous results predicting the yield curve and underlines how the FED itself uses information from the curve to determine targets. The snake shaped volatility curve, that plots volatility over maturity, is consistent with FED’s target policy and with announcement effects on short term maturity.

II FOMC decisions after 1994

FED’s communication policy dramatically changed after 1994: they decided to announce FOMC’s decisions just after the meeting and to schedule meetings regularly. Only few exceptions to this rule have been made ever since it was established and only in case of emergency situations.

III Yield curve Model with FOMC Decisions
It is a dynamic stochastic model that describes policy decisions of scheduled FOMC meetings as a function of some variables that describe the economy just before the meeting begins.
These variables are the federal funds target, the spread between the target and the short rate and its volatility, and a number of macroeconomic data available. Target’s moves depend then on a stochastic process.
The author includes a small probability that a target changes happens in a non scheduled meeting.
Policy Rule in modelled over a set of information available just before the meeting starts and built under the hypothesis that the objective of policy is to adjust against shocks. A pricing Kernel is used to solving for yields

IV Estimation
The main source of the data used in this paper comes from Datastream. The author uses series for LIBOR, target, and swap rates from January 1, 1994 to December 31, 1998. The source for FOMC schedule is the Board of Governors of the Federal Reserve.
A Stochastic Maximum Likelihood estimator is the tool for solving the density approximation, every technicality is part of the appendix.

V Estimation Results
Due to the shortness of the sample, the model fails to describe the probability of a target move with unconstrained volatility even if the derived policy rule turns out to be better than a Taylor rule in describing the target. Half life of macroeconomic shocks appears to be ≈ one year, it surges to 2,5 years with constant volatility. This long life of the shock causes policy inertia, inducing the FOMC to move the target with a higher probability after a move (autocorrelation). This result underlines how FOMC reacts to long term shocks and does not care about short term fluctuations.
Also the short rate is strongly correlated with FOMC meetings, even if some previous works, Dai-Singleton(2000), found a weak correlation. This confirms the good fitting of the model with the snake shaped volatility curve. The fitted parameters also describe well the qualitative decision of moving the target when fed in a discrete choice model.



You can follow us on Twitter  @fintlab and facebook