Impact of Federal Reserve Forward Guidance on Interest Rate Expectations

(1)

Faculty of Economics and Business

Impact of Federal Reserve Forward Guidance on

Interest Rate Expectations

BSc Economics & Finance

Author: Kevin Belegu

Student number: 10621075

Supervisor: Gabriele Ciminelli

Date: 15/07/2016, Amsterdam

(2)

Statement of Originality

This document is written by Kevin Belegu who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and

its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

This study investigates the effect of forward guidance on zero coupon treasury bond yields in the US using an event study regression analysis. We use the responses to the Survey of Primary Dealers conducted by New York Fed’s in order to construct a proxy of market expectations regarding the stance of monetary policy. We find that forward guidance statements issued by the Federal Open Market Committee successfully manage to lower interest rate expectations in the US. However, results suggest that forward guidance is less effective when counting for market expectations due to the anticipation of the monetary policy actions by the market participants. We measure the effects of forward guidance using a dataset of daily bond yields that goes back to 2004 and account for the FOMC statements of August 2011 and January, September, and December 2012.

(4)

The Federal Reserve has clear and specific goals it has committed to fulfill. It takes action through its monetary policy instruments in order to ensure stable inflation, economic growth and low unemployment rate. After the wake of the financial crises many central banks including the Federal Reserve started lowering their policy rates so as to stimulate the economy and promote recovery. However, once the effective federal fund rate reached the zero lower bound in 2008, it was not possible to further decrease the nominal short term rate. At this point the Federal Open Market Committee (hereafter: FOMC) was forced to move away from the traditional monetary policy and experiment with new unconventional methods to achieve policy accommodation.

The most popular unconventional tools were the large-scale asset purchases and forward guidance. This paper focuses on the latter and its effectiveness to influence the path of future interest rates. Over the past years, the FOMC has used its main communication tools like policy statements and press releases to convey its explicit policy rate guidance.

By implementing forward guidance, the FOMC reveals information to the public about the way that monetary policy is conducted and most importantly, about the future path of interest rates (Woodford, 2012). This type of communication narrows the information gap for the public and allows them to rely on newly formed expectations about the likely direction of monetary policy (Sheard, 2013). Business and investors reconsider their investment decisions based on the new future short-term rate expectations. According to the expectation theory, the interest rate of a long-term bond equals the average of the expected short term interest rates over the life of that same bond (Mishkin, Mathews, & Giuliodori, 2013). Based on this theory, it is hoped that by lowering future short-term rates forward guidance can also impact the path of medium to long-term interest rates and thus boost aggregate demand.

This study aims to quantify the impact of policy guidance on market expectations about the future path of interest rates. Our empirical analysis attempts to answer the following research questions:

Q1. Does forward guidance successfully affect the short to medium term zero coupon treasury yields in the US?

Q2. Did the market participants anticipate forward guidance statements?

In order to address these questions, we combine outcomes of the existing literature with our own empirical findings.

(6)

The first part of our empirical analysis focuses on the effect of forward guidance statements issued by the FOMC on zero coupon treasury bond yields with maturity ranging from one to ten years ahead. We implement an event study regression analysis where the first difference change in daily bond yields constitutes the dependent variable. The forward guidance dummy is the main explanatory variable.

Contrary to previous studies, in the second part of our analysis we incorporate a new explanatory variable in the regression that counts for monetary policy market expectations and thus captures only the surprise component of forward guidance statements. A proxy for market expectations is constructed based on the responses to the survey of primary dealers. This survey is conducted by the New York Fed two to three weeks before the FOMC releases its statement and it aims to capture investor’s views about monetary policy and the general economic outlook. The surprise component of forward guidance is calculated as the difference between the content of the actual statement release and the primary dealers expectations.

The remainder of the study is structured in the following way: Section 2 begins by providing arguments in favour and against forward guidance, followed by some historical remarks on the evolution of forward guidance statements in the US and an overview of the existing academic findings. Section 3 presents the dataset and describes the method that we use to assign discrete values to the responses of the survey of primary dealers and to the FOMC statement releases. Section 4 explain the methodology and section 5 provides a summary of the results. In section 6 we extend our empirical analysis and section 7 concludes.

2.Forward Guidance in Theory and Practice

The concept of policy guidance is not new for central banks around the world. Even before the burst of the financial crises, monetary policy committees revealed information regarding the future path of their policy rates. However, the policy guidance statements have been subject to changes throughout the years in order to better communicate central bank goals and objectives (den Haan, 2013). The first part of this section provides theoretical arguments in favour and against the use of forward guidance as an unconventional monetary policy tool. The second part describes the different types of forward guidance and its evolution through time in the United States. The last part provides an overview of the existing academic findings that are in line with our research questions.

(7)

2.1 Theoretical Arguments in Favour and Against Forward Guidance

Why does the Fed use forward guidance statements as an unconventional monetary policy tool? Are there any downside risks that can cause undesirable effects?

Goodhart (2013) argues that central banks have access to a broad range of information on the present and future economic conditions of their respective countries. Furthermore, central bank professionals and members of the monetary policy committee have adequate knowledge on the macroeconomic and financial developments. They should be able to predict the future course of monetary policy actions better than any other participant in the market. Not providing this information to the public would consist of an intentional withholding of beneficial information. Furthermore, den Haan (2013) argues that the implementation of forward guidance reduces the sensitivity of future interest rates and asset prices to surprise macroeconomic news releases. A lower sensitivity to macroeconomic data releases leads to lower volatility and thus improves the stability of financial market activities.

Forward guidance might be successful in decreasing the volatility of asset prices and interest rates, however Barwell and Chada (2013) raise the concern that this type of unconventional monetary policy is incomplete. After the first deviation of the interest rate from the zero lower bound, there is a lack of information about the economic circumstances and conditions that might affect the stance of monetary policy strategy. According to the authors, this information is important for investors to form their beliefs. These beliefs do matter because they have an impact on asset prices, future interest rates and economic activity today. By the time the first rate hike approaches, this lack of information could become a source of increased volatility and generate undesirable effects in the market (Barwell and Chada, 2013).

Williams (2013) points out some other issues concerning the effectiveness of forward guidance. Under weak economic performance the policy guidance statements enclose policy rate forecasts that stretch several years ahead. Fed credibility might be negatively affected by communications on the monetary policy actions so far off, especially when members of the FOMC might be different than the ones making decisions today. Additionally, Williams (2013) argues that the wording of forward guidance statements is of great importance in order to effectively convey the intended strategy to the public. Poor communication by the FOMC is likely to lead to disadvantageous effects and cause confusion in the market. Another problem concerning forward guidance is the risk of causing over-reliance on Fed statements. The public might tend to replace individual economic thinking and extensively focus on the content of Fed statements (Williams, 2013).

(8)

2.2 Categories of Forward Guidance and its Implementation in the US.

Campbell et al. (2012) are the first ones to distinguish between two different types of policy guidance, Odyssean and Delphic. Through the Delphic element of forward guidance, central banks provide a forecast of future macroeconomic performance and the likely monetary policy actions that can occur based on the policymakers’ superior information about macroeconomic fundamentals. This type of forward guidance does not publicly commit policymakers to particular actions in the future (Campbell et al., 2012). On the other hand, the Odyssean element of forward guidance does publicly commit policymakers to a course of action in the future. “Just as Odysseus committed to tie himself to the mast so he could not jump in the sea when hearing the sirens’ song, the central banks hold to a particular promised monetary strategy in the future” (Campbell et al., 2012).

Conditionality is another important component of forward guidance statements issued by central banks. Even though policymakers have superior information on the economic outlook, there is still some uncertainty about future developments which is reflected by the conditionality of their statements. Based on this, the ECB Monthly Bulletin (2014) distinguishes between four different categories of forward guidance:

1. Pure qualitative forward guidance. It does not provide any information about the end date or any numerical threshold related to the possible evolution of the policy rate in the future.

2. Qualitative forward guidance conditional on a narrative. This type of policy guidance provides qualitative statements about the possible evolution of policy rates, accompanied by a description of the macroeconomic circumstances under which the monetary policy might deviate.

3. Calendar based forward guidance. This type of statements reveal information about the likely path of policy rates up to an explicit date in the future. It is expected that the stance of monetary policy changes after this date.

4. Outcome based forward guidance provides explicit information about numerical conditions or thresholds that are linked to certain macroeconomic variables like inflation or unemployment rate. This information serves as a demarcation line. The public should infer that after the chosen variables meet their numerical requirements policymakers will reassess the best fitting course for the future federal fund rate.

Forward guidance statements implemented by the FOMC have experienced significant changes through their content and wording in the past years. Financial crises have forced policymakers to review their policy guidance statements so as to be more elaborate and

(9)

detailed. It was in February 2001 when the FOMC released for the first time formal and regular forward guidance by including a “balance of risks” in their statement. This was repeated again in August 2001. The FOMC was constantly experimenting with the wording of their statements in order to make their internal decision making clear and transparent to the public (Campbell, 2013).

In December 2008, the committee released a pure qualitative forward guidance statement implying that the federal fund rate was expected to remain low “for some time”. Even though this statement influenced public expectations about the future short term rates it was not exactly what the FOMC was expecting. During 2009-2011, the expectations from the financial market were not fully in line with the intended monetary policy. Despite the fact that the policymakers attempted to communicate the necessity for an extended period of highly accommodative policy, the market was expecting a lift-off of the federal fund rate in just a few quarters. This called for further action so as to change the tight policy expectations (Williams, 2013).

In August 2011, the FOMC shifted to what is known as a calendar based forward guidance and ensured the market that the federal rate will remain at exceptionally low levels at least through mid 2013. This statement facilitated the process of bringing market expectations in closer alignment with the intended monetary policy strategy. Furthermore, in December 2012, the committee introduced an outcome based forward guidance statement. The main goal was to improve public understanding of the Fed’s decision making process with respect to certain developments in specific economic variables. At this point the monetary policy was tied to changes in the unemployment rate. The FOMC set a numerical threshold of 6.5 percent and every level of unemployment below this threshold meant that the federal fund rate was likely to pick up so as to be consistent with the chosen monetary policy strategy at the time (Williams, 2013). In order to further support forward guidance effectiveness and reduce public confusion, all the FOMC members publish their own forecast for the optimal future path of the federal fund rate and other macroeconomic variables like inflation, unemployment and gross domestic product growth every three months.

2.3 Empirical Evidence

What is more important in influencing market expectations, Fed words or actions? In order to answer this question, we start by reviewing one of the most influential papers on forward guidance written by Gürkaynack, Sack and Swanson (2005). The authors begin by analyzing

(10)

the impact of changes in the federal funds rate change on asset prices and treasury yields. In order to quantify this impact, they use high frequency data for three different time windows. A short window (10 minutes previous to the announcement and 20 minutes after the announcement), a wider window (20 minutes previous to the announcement and 45 minutes after) and a daily window. These narrow time windows are important because they allow for an isolation of the impact of Fed actions from other macroeconomic news. The change in asset prices and treasury yields are linked to the surprise component of the federal funds rate by the following regression:

∆yt = α + β∆xt + εt

In this regression the change in asset prices and treasury yields, denoted by ∆yt,is the depended

variable and its variation is explained by the surprise component of the change in federal fund rate (∆xt) . The stochastic error term (εt) captures other factors that might impact treasury yields

and asset prices. The results suggested that the change in the federal funds rate had a significant impact on the S&P500 index and treasury yields of different maturities (Gürkaynack et al., 2005). Even though the results were significant, the regression was not able to distinguish between the factors that caused the surprise component. The authors did not reject the null hypothesis that more than 1 factor can explain for the variation in prices. Was it the change of the federal fund rate (target factor) or the policy guidance (path factor)? Since some of the FOMC statements provided information about changes in both the path and target factor simultaneously, Gürkaynack et al. (2005) distinguish between the two by using a factor analysis. The authors implement a principal component analysis, a procedure that converts correlated variables into uncorrelated variables by using an orthogonal transformation. The use of this method made possible to isolate the effects caused by a change in the federal fund rate and the change in forward guidance content. Interestingly, Gürkaynack et al. (2005) find that the path factor has a greater impact on long term treasury yields compared to the target factor, even when the change in the one year ahead rate is the same as the change in the federal fund rate. The path factor explained the variation in prices and yields three to ten times more than the target factor. These surprising results called for further forward guidance effectiveness.

Campbell et al. (2012) build their research on the findings of Gürkaynack et al. (2005). They extended the data sample and analyzed the effectiveness forward guidance statements for the period before and after the financial crises. First, they considered the announcements between February 1994 and June 2007, and after for the period August 2007 and December 2011. The pre-crises result show that the path factor explains 67 percent of the variation in the

(11)

expected fund rate for a period of two quarters ahead, and 90 percent for crises with the path factor explaining 53 percent of the variation in the expected fund rate two quarters ahead and 79 percent for six quarters ahead.

Woodford (2012) attempts to find further evidence on the effectiveness of forward guidance statements and their ability to influence market expectations. The author focuses on the evolution of the overnight interest rate swap contracts (hereafter: OIS) before and after the FOMC statement release. OIS are financial instruments that allow major banks to lend and borrow from each other at the overnight interest rate. Normally, if the monetary policy is being effective this should be reflected by changes in the overnight rate. Woodford (2012) tries to capture the immediate effect of the FOMC policy guidance by analyzing the movements in the overnight rate right before and after the announcements. The author uses high frequency transaction-by-transaction data and accounts only for the statements of August 2011 and January 2012, because they were not associated with any policy rate cuts or asset purchases. This way it becomes easier to isolate the effect of forward guidance statements on market expectations.

Figure 1: Intraday OIS rates on August 9th_{2011. Source:}_{Woodford (2012)}

Figure 1 shows the response of the OIS rates at four different maturities before and after the announcement of August 2011 which is represented by the dotted line. Every respective symbol depicted in the legend of figure 1 represents the overnight rate of a single OIS contract traded at that time. OIS contracts can have different rates depending on the byer and seller, even if

(12)

traded at the same time. This is why the representative symbols are not in a straight line and they intersect each other. It is clear that the rates experience a significant drop, specially the ones with a maturity of two and one and a half years ahead (Woodford, 2013). Very similar market reactions were observed after the announcement of January 2012 where the FOMC extended the assurance for low rates at least through mid-2013.

Figure 2: Intraday OIS rates on January 25th_{2012. Source:}_{Woodford (2012)}

As depicted from figure 2 the two year and one and a half overnight rates experienced a significant drop immediately after the announcement and than started picking up slowly one hour later, but still remained below the initial level (Woodford 2012). Different than the August 2011 statement, in December 2012 the overnight rates of the OIS contracts trading before the FOMC announcement started going down. This can be interpreted as a correct anticipation of the direction of monetary policy from the market participants.

Additionally, Swanson and Williams (2012) provide some interesting evidence on the effectiveness of forward guidance in influencing market expectations. They analyse the median response from the Blue Chip Forecast where market participants are asked to provide an estimate on the number of quarters until the FOMC policy rate will start picking up.

(13)

Figure3. Expected number of quarters until the first federal funds rate increase above 25 bp, from the monthly

Blue Chip survey of forecasters. Source: Swanson and Williams (2012)

In December 2008 when the policy rate was reduced to 0-25 basis points the FOMC announced that it was expecting the rate to remain at this level “for some time”. Figure 3 shows that the median response after this statement fluctuated within the range of two to five quarters ahead. In August 2011 when the FOMC switched to calendar based forward guidance by introducing the term “mid-2013” the median response shifted to seven or more quarters. By this time, market expectations were closely aligned with the FOMC’s own views.

These results showed a significant impact of forward guidance statements and suggest that market participants consider the FOMC forecast as superior and credible. To further support these findings, Swanson and Williams (2012) analysed the daily data of interest rate options with a maturity of five quarters ahead.

(14)

Figure 4: Probability the Federal Funds Rate would be less than 50 bp in five quarters. Source: Swanson and

Williams (2012)

They calculated the implied probability distribution of the federal fund rate remaining below 50 basis points in the next five quarters. As figure 4 depicts, after the forward guidance statement of August 2011, the probability distribution picks up and fluctuates between 80 and 90 percent on most of the days.

In a more recent study Filardo and Hofman (2014) analyze the impact of seven different forward guidance statements issued by the FOMC on three month Eurodollar future rates and treasury yields that mature in one, two, five and ten years. This statements were announced between December 2008 and December 2013.

Figure 5: Announcement effect on three-month interbank futures and bond yields for the Federal Reserve.

(15)

As it is shown in figure 5, the future rates follow a downward trend after most of the announcement days. The qualitative statements of December 2008 and March 2009 were the most effective in lowering long term treasury yields, however it is hard to say whether this was purely due to forward guidance because on these days there were also announcements of changes in the federal fund rate and large scale asset purchases (Filardo and Hoffman, 2014). The authors also point out that the market reaction to forward guidance statements decreased with time. For instance, the announcement of August 2011 led to a 20 basis points drop of the two year treasury yield, whether after the January and September 2012 announcements this drop was only 5 basis points (Filardo and Hoffman, 2014). December 2012 and 2013 were the only days were there was not a significant market reaction to forward guidance statements. However, Filardo and Hoffman (2013) argue that on these dates the effect was contaminated by the announcement that the Fed would slow down with its large scale asset purchases. In this case, the weak market reaction can be attributed to the effectiveness of forward guidance.

There are several other studies which attempt to assess the effectiveness of forward guidance statements by using dynamic stochastic general equilibrium models (DSGE). For instance, Smets and Wouters (2007) conclude that forward guidance leads to an increase of nominal treasury yields. On the other hand, Del Negro et al. (2014) use the DSGE model applied by the New York Fed and find that forward guidance reduces the long term treasury yields. Their study had more robust results compared to other structural models (De Graeve et al., 2014). Moessner (2015) investigates whether this robust prediction by the DSGE models is in fact consistent. The author runs an event-study regression with changes in nominal treasury yields as a dependent variable and a dummy variable that equals 1 on the days of the announcements and 0 otherwise. Moessner (2015) controls for 11 different macroeconomic variables in order to isolate the effect of forward guidance on long term treasury yield. The author concludes that for the forward guidance statements of December 2008, March 2009, August 2011, January, September and December 2012 the treasury yields at maturities of two to five years ahead experienced a significant drop.

To conclude, the existing literature suggests potential effectiveness of forward guidance in affecting medium to long term treasury yields, however there is still an ongoing debate about the consistency and efficiency of this unconventional monetary policy tool at the zero lower bound. In this paper we further extend the study of Moessner (2015) by completing the event study regression with another market expectation variable. The following sections explain the

(16)

way we proceed with our data selection, market expectation variable construction and regression analysis.

3. Data

In line with Moessner (2015) this study investigates the reactions of the nominal zero coupon treasury yield of maturities from one to ten years ahead computed following the methodology of Gürkaynack et al. (2006). The zero coupon treasury yields are used because of their high sensitivity to the release of monetary policy announcements or other type of macroeconomic news. The data is obtained from the Fed website and the time window goes from January 2004 to June 2016. We control for the effect of other news on treasury yields by including in our regression the surprise component of ten different macroeconomic variables that have a significant impact on zero coupon treasury yields, in line with the study of Faust et al. (2007) and Balduzzi et al. (2001). These macroeconomic variables are: Capacity Utilization, Housing Starts, Producer Price Index (PPI), New Home Sales, Consumer Confidence, Consumer Price Index (CPI), Nonfarm Payrolls (NFP), Initial Claims, Unemployment, and Durable Goods Order. Data is obtained from datastream and the surprise component is calculated as the difference between the actual macroeconomic release and the median Reuters Poll forecast. The forward guidance statements of August 2011 and January, September, and December 2012 are taken into account. We focused on these four statements because they contain a surprise wording compared to the previous statement released by the FOMC and thus are likely to generate a more significant market reaction. Additionally, we construct two new market expectation variables about forward guidance and asset purchases based on the responses to the survey of primary dealers (hereafter: SPD). This survey is conducted by New York’s Fed and it aims to understand expectations of market participants about the likely stance of monetary policy. The following section explains how we use the results from the SPD in order to construct the new market expectation variables.

3.1 Forward Guidance Surprise Variable

The responses to the survey of primary dealers conducted by New York Fed are very important for Fed’s understanding of market expectations regarding monetary policy. The survey takes place two to three weeks before the next FOMC meeting and the results are released three weeks after this meeting. The SPD is currently sent out to the main 21 institutions involved in trading government and other type of securities with the Fed. These 21 institutions represent the primary dealer community and their market expectations are of crucial importance because

(17)

they have a significant number of clients. So, in a way, primary dealer monetary policy expectations reveal information about investor beliefs and thus can serve as a market expectation proxy (Golay et.al, 2013). The SPD consist of two main parts, the first one concentrates on the expectations regarding monetary policy and the second part on economic outlook. In this study we focus mostly on the first part of the SPD, more precisely on the second question of the first part. This questions asks the the recipients whether they expect any changes in the next statement of the FOMC and, if so, what do they expect to change? Recipients can answer the question freely in their own words, without any limitations.

After reviewing the SPD responses we assign numerical values to all of them as follows:

Very hawkish forward guidance expectations -2

Hawkish forward guidance expectations -1

Neutral forward guidance expectations 0

Dovish forward guidance expectations +1

Very dovish forward guidance expectations +2

Where hawkish suggest monetary policy tightening and dovish suggest monetary policy easing. We follow the same method in order to assign numerical values to the actual FOMC statement releases based on the content of forward guidance. However, it should be pointed out that the actual statements are always compared to the content of the previous statement releases. For example, suppose it is June 2015 and in the last statement the FOMC assured the market that interest rates will remain low at least until mid-2016. The June forward guidance statement is considered dovish (easing) only if it assures the market that the policy rate is going to remain low for a period longer than mid-2016, otherwise it is either a neutral or hawkish statement. After comparing the actual statement to the previous ones we assign numerical values as follows: Very hawkish forward guidance statement -2

Hawkish forward guidance statement -1

Neutral forward guidance statement 0

Dovish forward guidance statement +1

Very dovish forward guidance statement +2

Again, hawkish suggest tightening monetary policy and dovish easing monetary policy. After assigning values to the actual statement releases and the SPD responses we calculate the surprise component as follows:

(18)

Surprise Forward Guidance = Actual Statement Release – SPD Response

Based on the values of the actual statement release and SPD responses the variable capturing the surprise component of forward guidance can take the discrete values of -4, -3, -2 , -1, 0, 1, 2, 3, and 4.

For instance, in the statement of August 2011 the FOMC changed the policy guidance wording from “likely to warrant exceptionally low levels of the federal fund rate for an extended period” in July to “[…] likely to warrant exceptionally low levels of the federal fund rate for at least through mid-2013”. This was the first time the FOMC was committing to maintain low rates up to a specific period in the future. The calendar based forward guidance intended to send a clear message of monetary policy easing to, thus we assign the value of 2 to the actual statement release. On the other hand, the 21 primary dealers were not expecting any substantial changes in the FOMC policy guidance statement.

“Some dealers did not expect any substantial changes to the statement” (New York Fed, Responses to Survey of Primary Dealers, August 2011)

This was also shown in the median response of the primary dealers for the likely federal fund rate in the second quarter of 2013, equaling 1 percent and suggesting that a forward guidance extension was not expected. Since no change was expected we assign the value of 0 to the SPD response.

In January 2012 the FOMC revised its forward rate guidance compared to the statement of December 2011. The “[…] exceptionally low levels of the federal funds rate at least through mid-2013” phrase was dropped and substituted by “[…] exceptionally low levels of the federal funds rate at least through late 2014”. Extending the assurance for low rates is a clear policy easing message. We consider this as a very dovish forward guidance statement and assign the value of 2 to the actual statement release. Different than in August 2011, the SPD response of the January 2012 meeting the shows that primary dealers this time were partially anticipating some changes in the FOMC statement.

“Several dealers believed the committee would drop the phrase “at least though mid-2013” form the statement entirely, while some predicted that the calendar based forward guidance would remain, but be pushed out further into the

(19)

Since several dealers were expecting changes with only some of them expecting a further easing of monetary policy we assign the value of 1 to the SPD response.

In the statement of September 2012 the FOMC pushed further into the future the low rate policy guidance compared to the statement of August 2012. The calendar based forward guidance “[…] exceptionally low rates at least through late 2014” was substituted by “[…] exceptionally low rates at least through mid-2015”. The FOMC once again committed to maintain low rates up to a later date in the future. This was a clear policy easing communication in order to bring market expectations in line with the committee’s decision making. Thus we assign the value of 2 to the actual statement release. Interestingly, the SPD response for this statement shows that market expectations were in line with the committee policy actions.

“Most dealers expected some sort of easing to be announced in the September FOMC statement. Most dealers expected that the forward guidance on the path of the federal funds rate would be extended into 2015. Several dealers specified that the guidance to be extended to “mid-2015” and couple of dealers expected that the forward guidance could be extended past 2015” (New York Fed,

Responses to Survey of Primary Dealers, September 2012)

Since most of the primary dealers correctly anticipated the actions of the FOMC regarding the policy guidance of future interest rates, we assign the same value of two to the SPD response. In December 2012 the FOMC switched to outcome based forward guidance. The phrase “[...] at least through mid-2015” was replaced by the 6.5 percent unemployment threshold. Responses of the SPD showed that the market participants were expecting the unemployment rate to fall below the 6.5 percent threshold only after the last quarter of 2015, suggesting that the low rate policy guidance was extended even further when compared to the previous mid-2015 qualitative statement. However, since the future path of the unemployment rate is subject to unknown fluctuations and can be misinterpreted by the market participants, this statement is considered as dovish and we assign the value of 1 to the actual statement release. On the other hand, SPD responses of December 2012 did not mention any expected changes regarding forward guidance in the next FOMC announcement. Furthermore, the median response of the likelihood that the FOMC will adopt qualitative or quantitative thresholds to provide policy guidance was only 10 percent, suggesting again that no significant change was expected in the

(20)

stance of forward guidance. Thus we assign the value of 0 to the SPD response. Overall, after calculating the surprise forward guidance variable for each of the four FOMC statements that we account for we obtain the following values: +2, +1, 0, and +1.

3.2 Asset Purchase Surprise Variable

The FOMC provided information about it’s asset purchase program only on the announcements of September and December 2012. However, after reviewing the SPD responses we conclude that on both of these dates the primary dealers correctly anticipated the decisions of the committee regarding asset purchases.

“Some dealers noted that the announcement of an asset purchase program was

possible, with a few dealers noting the possibility of open-ended purchases.” (New

York Fed, Responses to Survey of Primary Dealers, September 2012)

and

“Nearly all dealers noted that they expected the FOMC to announce the continued purchase of Treasury securities following the conclusion of the Maturity Extension

Program, at the December meeting.” (New York Fed, Responses to Survey of Primary Dealers, December 2012)

In the August 2011 and January 2012 statements the FOMC did not reveal any information about the stance of asset purchases. At the same time, the respective SPD responses show that the primary dealers were also not expecting any purchases by that time. After these observations, if we calculate the surprise component of the asset purchase variable we obtain only zero values.

4.Methodology

The empirical research of this paper has two parts. The first part focuses on the effect of forward guidance in influencing short to medium term zero coupon treasury yields. The second part attempts to find evidence whether the market participants did in fact anticipate the forward guidance statements.

(21)

4.1 Event Study Regression Analysis

This study implements the methodology of Moessner (2015) in order to quantify the reactions of treasury yields after the FOMC policy guidance statements. We begin by calculating the first difference in daily zero coupon bond yields of maturities ranging from one to ten years ahead.

Δy(t) = [y(t) – y (t-1)] (1)

The daily changes in treasury yields, Δy(t), constitute the dependent variable which is regressed on a forward guidance dummy variable. We also control for the surprise component of the ten macroeconomic news releases in order to isolate the effect of the FOMC policy guidance statements from other factors that might also have an impact on treasury yields. Our basic regression looks as follows:

Δy(t) = c + a * d_FG(t) + ∑ (𝑏𝑖∗ 𝑠𝑢𝑟𝑝𝑟𝑖𝑠𝑒𝑖(𝑡)) 10

𝑖=1 + ε(t) (2)

In the above regression the dummy variable d_FG(t)equals one on the days of the FOMC

announcements and zero otherwise. The ∑10 (𝑏𝑖∗ 𝑠𝑢𝑟𝑝𝑟𝑖𝑠𝑒𝑖(𝑡))

𝑖=1 , i = 1,…,10, term captures

the effect of the ten macroeconomic data releases on daily treasury yields.. We measure the effect of forward guidance by running ten regressions for bond yields with a maturity of one to ten years ahead. The advantage of this event study regression is its simplicity and parsimony (Moessner, 2015). On the other hand, within this approach it is not possible to count for dynamic interactions between yields of different maturities, neither for correlation of error terms.

We expect coefficient a, which represents the change in daily yields to be significantly lower than zero in regression (2). This in fact would suggest the use of policy guidance from the FOMC is successful in steering interest rate market expectations.

4.2 Market expectations

In the second part of our empirical analysis we extend the event study regression of Moessner (2015) by including the variables that were constructed in section three. These variables capture the surprise component of the FOMC statements. In other words, the difference between the

(22)

actual statements and the market expectations of the primary dealers. We denote the surprise variable by S_FG for forward guidance and S_AP for asset purchases. However, as explained

in section three, after constructing the variable that captures the surprise component of asset purchases we obtain only zero values due to the correct anticipation of the monetary policy action by the market participants. This means that even if this variable is used to extend regression (2) it will not affect the results and thus it is not included. The new regression takes the following form:

Δy(t) = c + a1 * S_FG(t) +∑ (𝑏𝑖∗ 𝑠𝑢𝑟𝑝𝑟𝑖𝑠𝑒𝑖(𝑡)) 10

𝑖=1 + ε(t) (3)

In regression (3) the variable S_FG(t) accounts only for the surprise component of forward

guidance. In other words, the better market participants anticipate forward guidance statements the more S_FG value approaches 0. Therefore, we expect the absolute value of coefficient a1 in

regression (3) to be higher than the absolute value of coefficient a in regression (2). In order to better understand this hypothesis, consider the following example: assume that after the announcement of August 2011 the five year zero coupon treasury yield fell by 10 basis points. If we ignore the effect of the control variables than under regression (2) this change is explained by the value of the dummy (d_FG = 1)times the respective coefficient a. Under regression three

the 10 basis point change should equal the product of the S_FG value and coefficient a1. If the

value S_FG approaches 0 (market anticipates forward guidance) than the absolute value of a1

should increase in order to compensate for the 10 basis point drop.

Additionally, we extend our empirical analysis by repeating regressions (2) and (3), but now for a wider time window of three days. A wider window allows to check whether the effects of the variables d_FG and S_FG on zero coupon treasury yields are persistent over time

and thus economically meaningful.

5.0 Results

This section provides an overview of the results that we obtained throughout our empirical analysis. We begin by analyzing the results of regression (2). Table 1 summarizes the results of ten different regressions where the depended variables are daily changes in zero coupon treasury yields of ten different maturities. In line with our expectations the regression

(23)

coefficients for the forward guidance dummy have a negative sign, indicating that the FOMC announcements led to a reduction in zero coupon treasury yields. These negative coefficients have a significant impact on treasury yields at the 1 % level for maturities of five to nine years in the future. The significance level increases to 5% for maturities of four and ten years ahead in the future. The coefficient a is significant at the 10% level only for treasury yields maturing in three years. Whether for the yields with a maturity of one and two years in the future, the results suggest that forward guidance did not have a significant impact. The yields with a maturity of six and seven years ahead experienced the largest reduction of 8 basis points.

Furthermore, the adjusted R-squared value of regression 2 ranges from 0.0263 to 0.0416 for all the different maturities. This relatively low R-squared is due to the fact that our dependent variables are daily changes in treasury yields and for many days there are no explanatory variables to explain the variability of zero coupon treasury yields (control variables are in approximately monthly entries). Even though our data sample is extended with an additional three years, these findings are in line with those of Moessner (2015).

Overall, the results obtained for the first part of our empirical analysis support the hypothesis that the implementation of forward guidance as an unconventional policy tool by the FOMC is able to significantly lower interest rate market expectations at the horizons of three to ten years ahead.

Table 2 shows the results of regression (3) which includes the surprise forward guidance variable (S_FG) constructed in section three. The results show that the surprise

component of the forward guidance statements that was not anticipated from the market participants has a significant impact on zero coupon treasury yields. The negative coefficient a1 is significant at the 1% level for yields maturing in four to ten years ahead. This significance

level increases at 5 % for yields maturing in three years and to 10% for yields maturing in two years. The coefficient a1 is not significant only for yields maturing in one year.

In line with our expectations the coefficient a1 from regression (3) has a higher absolute

value than the coefficient a of regression (2) at the horizons of five to ten years ahead. The coefficient a1 remains lower than a for horizons of one to four years which can be explained

by the different values that we assign to our explanatory variables in regression (2) and (3). These results support the hypothesis that market participants were able to correctly anticipate the direction of monetary policy. Additionally, by looking at the Adjusted-R-Squared values of regression (3) we can realize that they are higher than those under regression (2) at all time horizons, suggesting that the variable that captures market expectations (S_FG) is a better fit to

(24)

To conclude, the FOMC forward guidance announcements of August 2011, January 2012, September 2012, and December 2012 were followed by market reactions that led to a significant reduction of medium to short term future interest rates, thus effectively bringing market expectations closer to the committee’s intended policy strategy. However, results of regression (3) provide evidence that market participants were able to anticipate the content of forward guidance statements. The FOMC should closely consider investor’s views about the stance of monetary policy when formulating the content of it’s forward guidance statements. Adapting statement content to market expectations should facilitate the process of achieving a market reaction that supports the FOMC goals and objectives.

(25)

Table 1. Results of regression (2). The values in parenthesis represent the respective standard errors. Coefficients followed by one, two, and three stars are significantly different than 0 at 10%, 5%, and 1% level, respectively. Sample period : 01 January 2004 - 01 June 2016

Variable / Maturity in Years 1 2 3 4 5 6 7 8 9 10

d_FG -0.0119 -0.0338 -0.0527* -0.0669** -0.0761*** -0.0805*** -0.0807*** -0.0779*** -0.0726*** -0.0657**

(0.019) (0.024) (0.026) (0.028) (0.029) (0.029) (0.026) (0.029) (0.029) (0.029)

DURABLE GOODS ORDER -0.0008 -0.0013 -0.0015 -0.0015 -0.0014 -0.0014 -0.0013 -0.0014 -0.0014 -0.0015

(0.001) (0.001) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) UNEMPLOYMENT -0.0379 -0.0497 -0.0518 -0.0502 -0.0481 -0.0467 -0.0461 -0.0461 -0.0463 -0.0466 (0.022) (0.027) (0.030) (0.032) (0.033) (0.033) (0.033) (0.033) (0.033) (0.033) INITIAL CLAIMS -0.0006 -0.0001 -0.0001 -0.0001 -0.0001 -0.004 0.00005 0.00004 0.0008 0.0001 (0.0001) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) NONFARM PAYROLL 0.0004*** 0.0006*** 0.0007*** 0.0007*** 0.0007*** 0.0007*** 0.0006*** 0.0006*** 0.0006*** 0.0006*** (0.0004) (0.00005) (0.00006) (0.00006) (0.0006) (0.0005) (0.0007) (0.0007) (0.0007) (0.0007) CPI 0.0178 0.0362 0.0499 0.0576 0.0608 0.0609 0.0591 0.05635 0.0530 0.0497 (0.035) (0.043) (0.048) (0.051) (0.052) (0.053) (0.053) (0.0053) (0.053) (0.053)

NEW HOME SALES -0.0653 -0.0310 -0.0129 0.0057 0.0272 0.0497 0.0704 0.0885 0.103 0.1141

(0.106) (0.130) (0.145) (0.153) (0.157) (0.159) (0.161) (0.160) (0.161) (0.161) CONSUM.CONDFIDENCE 0.00067 0.0009 0.0012 0.0013 0.0014 0.0015 0.00165* 0.0017* 0.0018* 0.0018* (0.006) (0.0007) (0.0008) (0.0009) (0.0009) (0.0009) (0.0009) (0.0009) (0.0009) (0.0009) PPI 0.0147** 0.0133 0.0155 0.0183 0.0208* 0.0228** 0.0243** 0.0254** 0.0263** 0.0269** (0.007) (0.008) (0.010) (0.010) (0.010) (0.011) (0.011) (0.0009) (0.011) (0.011) HOUSING STARTS 0.0848** 0.1142** 0.1295** 0.1355*** 0.1354** 0.1312** 0.1247** 0.1171* 0.1092* 0.1016* (0.040) (0.049) (0.054) (0.057) (0.059) (0.060) (0.060) (0.050) (0.060) (0.060) CAPACITY UTILIZATION 0.0245*** 0.03402*** 0.0360*** 0.0334*** 0.0287** 0.0233* 0.0183 0.0141 0.0107 0.0082 (0.008) (0.0120) (0.054) (0.012) (0.013) (0.013) (0.013) (0.012) (0.013) (0.013) CONSTANT -0.00008 -0.00007 -0.0001 -0.0002 -0.0001 -0.0003 -0.0003 -0.0004 -0.0005 -0.0006 (0.0007) (0.0008) (0.0120) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) R SQUARED 0.0298 0.0432 0.0450 0.0436 0.0412 0.0387 0.0364 0.0342 0.0323 0.0307 ADJUSTED R SQUARED 0.0263 0.0398 0.0416 0.0402 0.0378 0.0353 0.0329 0.0308 0.0289 0.0272 NO. OF OBSERVATIONS 3118 3118 3118 3118 3118 3118 3118 3118 3118 3118

(26)

Table 2. Results of regression (3). The values in parenthesis represent the respective standard errors. Coefficients followed by one, two, and three stars are significantly

different than 0 at 10%, 5%, and 1% level, respectively. Sample period: 01 January 2004 – 01 June 2016

Variable\Maturity in Years 1 2 3 4 5 6 7 8 9 10

S_FG -0.0125 -0.0335* -0.0522** -0.0660*** -0.0768*** -0.0809*** -0.0805*** -0.0786*** -0.0752*** -0.0705***

(0.016) (0.019) (0.021) (0.023) (0.023) (0.024) (0.024) (0.024) (0.024) (0.024)

DURABLE GOODS ORDER -0.0008 -0.0013 -0.0015 -0.0015 -0.0014 -0.0014 -0.0013 -0.0014 -0.0014 -0.0015

(0.001) (0.0018) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) UNEMPLOYMENT -0.0379* -0.0497* -0.0518* -0.0502 -0.0480 -0.0467 -0.0461 -0.0461 -0.0463 -0.0466 (0.022) (0.027) (0.030) (0.032) (0.033) (0.033) (0.033) (0.033) (0.033) (0.033) INITIAL CLAIMS -0.0006 -0.0001 -0.0001 -0.0001 -0.0001 -0.00004 -0.0005 -0.00004 0.00008 0.0001 (0.0001) (0.0002) (0.002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) (0.0002) NONFARM PAYROLL 0.0004*** 0.0006*** 0.0007*** 0.0007*** 0.0007*** 0.0007*** 0.0006*** 0.0006*** 0.0006*** 0.0006*** (0.0004) (0.00005) (0.0006) (0.00006) (0.00006) (0.00007) (0.00007) (0.00007) (0.00007) (0.00007) CPI 0.0178 0.0362 0.0499 0.0576 0.0608 0.0609 0.0591 0.0563 0.0537 0.0497 (0.035) (0.043) (0.048) (0.051) (0.052) (0.053) (0.053) (0.053) (0.0537) (0.053)

NEW HOME SALES -0.0653 -0.0310 -0.0129 0.0057 0.0272 0.0497 0.0704 0.0885 0.0018* 0.1141

(0.106) (0.130) (0.145) (0.153) (0.157) (0.159) (0.160) (0.161) (0.1614) (0.161) CONSUM.CONDFIDENCE 0.0006 0.0009 0.0012 0.0013 0.0014 0.0015 0.0016* 0.0017* 0.0247** 0.0018* (0.0006) (0.0007) (0.0008) (0.0009) (0.0009) (0.0009) (0.0009) (0.0009) (0.0009) (0.0009) PPI 0.0144** 0.0126 0.0144 0.0169 0.0193* 0.0211** 0.0226** 0.0238** 0.0247** 0.0255** (0.007) (0.008) (0.009) (0.010) (0.010) (0.0009) (0.011) (0.011) (0.0111) (0.011) HOUSING STARTS 0.0848** 0.1141** 0.1293** 0.1353** 0.1351** 0.1310** 0.1244** 0.1169* 0.1090* 0.1014* (0.040) (0.049) (0.054) (0.057) (0.059) (0.060) (0.060) (0.060) (0.0608) (0.060) CAPACITY UTILIZATION 0.0245*** '0.0340*** 0.0360*** 0.0334*** 0.0287** 0.0233* 0.0183 0.0141 0.0107 0.0082 (0.008) (0.010) (0.012) (0.012) (0.013) (0.013) (0.013) (0.013) (0.013) (0.013) CONSTANT -0.0008 -0.00007 -0.0001 -0.0002 -0.0002 -0.0003 -0.0003 -0.0004 -0.0005 -0.0006 (0.0007) (0.0008) (0.0009) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) R SQUARED 0.0298 0.0435 0.0455 0.0443 0.0421 0.0398 0.0374 0.0353 0.0334 0.0318 ADJUSTED R SQUARED 0.0264 0.0435 0.0455 0.0410 0.0387 0.0364 0.0340 0.0319 0.0300 0.0284 NO. OF OBSERVATIONS 3118 3118 3118 3118 3118 3118 3118 3118 3118 3118

(27)

6. Extended Empirical Analysis

The dependent variables of regressions (2) and (3) are the daily changes of zero coupon treasury yields of ten different maturities. Using this short one-day window allows to control for the effect of other factors that might impact yields, due to the narrow time period. However, such a short period has the disadvantage that it might capture effects that do not hold for longer time windows, or it might just not capture the whole effect. Therefore, we modify the dependent variable by increasing the time window of yield changes from one to three days, replacing Δy(t) in regression (2) and (3) with the following:

Δ3_{y(t) = [y(t+2) – y(t-1)]}

Increasing the time window allows to check if the results that were obtained in the previous section were persistent over time and thus economically meaningful, or if they were reversed after the one-day period.

Table 3: Results of regression (2) for a wider time window of 3 days. Coefficients on the 10

macroeconomic control variables are not shown. Sample period: January 2004 – June 2016

Table 3 shows the results of regression (2) after extending the time window to three days. The negative coefficients representing the effect of the FOMC forward guidance statements are somewhat higher and they remain significant for yields at the horizons of 3-8 years ahead. This

Maturity in years d_FG Adjusted R2 No. of Observations

1 -0.0133 0.0212 3116 2 -0.0419 0.0298 3116 3 -0.0658* 0.0302 3116 4 -0.0839** 0.0287 3116 5 -0.0958** 0.0269 3116 6 -0.10223** 0.0251 3116 7 -0.1038** 0.0235 3116 8 -0.1018* 0.0220 3116 9 -0.0969 0.0207 3116 10 -0.0901 0.0196 3116

(28)

means that the forward guidance impact in lowering zero coupon treasury yields is not transitory after the one-day window but it persists for a period as long as three days.

We proceed in the same way also for regression (3). The coefficient a1 remains

significant at the horizons of 4-10 years ahead.

Table 4: Results of regression (3) for a wider time window of 3 days. Coefficients on the 10

Macroeconomic control variables are not shown. Sample period: January 2004 – June 2016

Interestingly the values represented in table 4 suggest that when allowing for a wider time window there is a considerable difference if we proxy for market expectations. The absolute values of the coefficient a1 are significantly higher for the three-day window. A possible

interpretation for this might be the difference in expectations between traders and primary dealers. Because traders have access to less “insider information” about the Fed’s actions compared to primary dealers they might overreact in the short run. However, in the longer-run the primary dealers are the ones that determine bond prices and the overreaction of traders fades away. The effects of this overreaction would persist only in case the primary dealers did not correctly anticipate the forward guidance statements, however in order to draw better conclusion on this matter it would have been better if we had access to intra-day data.

7. Conclusion

In this study we focus on the effectiveness of forward guidance in lowering short to medium term interest rates at the zero lower bound in the US. We also take into consideration market

Maturity in years S_FG Adjusted R2 No. of Observations

1 -0.0077 0.0264 3116 2 -0.0396 0.0400 3116 3 -0.0680* 0.0421 3116 4 -0.0913*** 0.0410 3116 5 -0.1089*** 0.0387 3116 6 -0.1210*** 0.0364 3116 7 -0.1284*** 0.0340 3116 8 -0.1318*** 0.0319 3116 9 -0.1321** 0.0300 3116 10 -0.1302** 0.0284 3116

(29)

expectations and their link to forward guidance effectiveness. Initially we provide some theoretical arguments in favour and against forward guidance and its evolution in the US. Later on, we attempt to answer our research questions by providing an overview of the existing academic findings, followed by an event study regression analysis based on the methodology of Moessner (2015).

The existing academic work suggests that interest rate expectations are sensitive to central bank communications just as much as the actions it takes on the policy rate. Most of the findings conclude that forward guidance can effectively lower market expectations on short to medium term interest rates in the US. However, it is important to notice that not all the FOMC forward guidance statements generated the desired outcomes. Furthermore, the use of forward guidance appears to be a good tool in narrowing the gap between the market expectations about the stance of the monetary policy and Fed’s reaction function.

The empirical analysis of the present study is conducted in two parts. Our first model incorporates a dummy variable equaling one on days that the FOMC released a policy guidance statement and zero otherwise. Results provide strong evidence that the use of forward guidance is successful in lowering interest rate expectations. The dummy variable that captures this effect is significant at the 1% or 5% level for zero coupon treasury yields maturing in four to ten years ahead. The results are robust to a wider time window (from one to three days) for maturities ranging from four to seven years ahead, in line with the findings of Moessner (2015).

Different to earlier contributions, in the second part of our empirical analysis we take into account market expectations regarding the stance of monetary policy. We incorporate a new variable that captures only the unanticipated (surprise) component of forward guidance statements in our model. The new variable is significant at the 1% level for yields maturing in four to ten years. The absolute value of the coefficient for this variable is higher when compared coefficient of the forward guidance dummy, providing evidence that market participants are able to correctly anticipate the FOMC monetary policy actions.

The use of forward guidance as an unconventional monetary policy tool by the FOMC seems to be effective in lowering interest rates expectations, however policymakers should pay close attention to market expectations when making decisions on the future course of monetary policy and specifically on the content of the forward guidance statements.

(30)

References

Balduzzi, P., Elton, E. J., & Green, T. C. (2001). Economic news and bond prices: Evidence from the U.S. Treasury market. The Journal of Financial and Quantitative Analysis,

36(4), 523.

Barwell, R., & Chada J. S. (2013). Complete forward guidance. Forward Guidance:

Perspectives from Central Bankers, Scholars and Market Participants. Vox eBook,

125 – 133.

Campbell, J. R., Evans, C. L., Fisher, J. D., Justiniano, A., Calomiris, C. W. & Woodford, M. (2012). Macroeconomic Effects of Federal Reserve Forward Guidance. Brooking

Papers on Economic Activity, 1-80.

De Graeve, F., Ilbas, P. and Wouters, R. (2014) Forward Guidance and Long Term Interest Rates: Inspecting the Mechanism. National Bank of Belgium, Brussels

Del Negro, M., Giannoni, M. and Patterson, C. (2012). The Forward Guidance Puzzle,

Federal Reserve Bank of New York Staff Reports No.574

den Haan, W. (2013). Forward Guidance: Perspectives from Central Bankers, Scholars and Market Participants. Center for Economic Policy Research, Vox eBook, 1-25.

Faust, J., Rogers, J. H., Wang, S. B., & Wright, J. H. (2007). The high-frequency response of exchange rates and interest rates to macroeconomic announcements. Journal of

Monetary Economics, 54(4), 1051-1068.

Federal Reserve Bank of New York (2011). Responses to survey of primary dealers.

Retrieved from https://www.newyorkfed.org/markets/primarydealer_survey.html Federal Reserve Bank of New York (2012). Responses to survey of primary dealers.

Retrieved from https://www.newyorkfed.org/markets/primarydealer_survey.html Filardo, A., Hofmann, B. (2014). Forward Guidance at the Zero Lower Bound, BIS Quarterly

Review.

Golay, E. C., Friedman, S. and McMorrow, M. (2013). Understanding the New York Fed’s Survey of Primary Dealers. Current issues in economics and finance, 19 (6).

Goodhart, C. (2013). Debating the Merits of Forward Guidance. Forward Guidance:

151-157.

Gürkaynak, R. S., Sack, B. & Swanson, E. T. (2005). Do actions speak louder than words? The response of asset prices to monetary policy actions and statements. International

Journal of Central Banking, 1, 55-93.

Gürkaynak, R. S., Sack, B., & Wright, J. H. (2006). The U.S. Treasury yield curve: 1961 to the present. Journal of Monetary Economics, 54(8), 2291-2304.

Kool, C. J., & Thornton, D. L. (2012). How effective is central bank forward guidance?.

Federal Reserve Bank of St. Louis Working Paper Series, 2012-063A.

Mishkin, F., Matthews, K., & Giuliodori, M. (2013). The Economics of Money, Banking, and

Financial Markets (European Edition). Harlow: Pearson Education.

Moessner, R. (2015). Reactions of real yields and inflation expectations to forward guidance in the United States, Applied Economics, 57:26, 2671-2682.

Sheard, P. (2013). A Central Bank Watcher’s Prespective. Forward Guidance:

157-165.

Smets, F. and Wouters, R. (2007). Shocks and Frictions in US Business Cycles: a Baesian DSGE Approach, American Economic Review, 97, 586-606.

(31)

Swanson, E. T., & Williams, J. C. (2012). Measuring the Effect of the Zero Lower Bound on Medium-and Longer-Term Interest Rates. Federal Reserve Bank of San Francisco

Working Paper, 2.

“The ECB’s Forward Guidance, Monthly Bulletin, ECB, Frankfurt am Main, April 2014. Williams, C. J. (2013). Forward Policy Guidance at the Federal Reserve. Forward Guidance:

43-49.

Woodford, M. (2012). |Methods of Policy Accommodation at the Interest-Rate Lower Bound, The Changing Policy Landscape: 2012 Jack Hole Symposium.

Impact of Federal Reserve Forward Guidance on Interest Rate Expectations

Faculty of Economics and Business