Printer Friendly

Is the fed funds rate still effective?

Both the depth and length of the Great Recession create the impression that the economy proved impervious to monetary policy. Policy rates, such as the federal funds rate (fed funds rate), were set at record lows, but the recovery in housing, employment, and GDP were subpar at best. Now with a self-sustaining expansion, the FOMC began rolling back its asset purchases program and, at some point in the future, it will start increasing its target for the fed funds rate. This raises the questions of whether the fed funds rate remains an effective tool and what effect an altered Federal Reserve balance sheet will have on inflation and the unemployment rate in the post-Great Recession world. Our econometric analysis suggests that since the 1990s, the traditional tools of monetary policy (such as the fed funds rate) may have influenced the unemployment rate but that it did not influence inflation. Thus, the effect a change in the fed finds rate may not be as straightforward as suggested by the conventional economic theory and the traditional link between interest rates and inflation and unemployment may have broken down.

Business Economics (2014) 49, 253-262.

doi:10.1057The.2014.32

Keywords: fed funds rate, inflation, unemployment rate, impulse response

Monetary policy does not operate in a vacuum. Context matters. The depth and duration of the Great Recession and the slow pace of recovery suggest that economic activity during this period was less responsive to traditional monetary policy than before the crisis. Has the link between the federal funds rate (fed funds rate) and the Federal Reserve's two targets for price stability and maximum employment changed?

Recently, the Federal Open Market Committee (FOMC) began rolling back its third asset purchase program, or QE3. At some point in the future, the FOMC will start increasing its target for the fed funds rate. This raises questions. What would be the likely effect of an increase in interest rates on inflation and the unemployment rate in the post-Great Recession world, and is that link different than before the Great Recession? In theory, interest rates, inflation, and the unemployment rate are closely related. However, economies and the relationships between economic variables evolve over time. Therefore, it would be useful for effective decision-making to quantify the precise statistical relationship between the fed funds rate, inflation, and unemployment rate in the post-Great Recession era. Specifically, do the data support a causal relationship between these variables in the post-Great Recession period? What would be the likely effect of a temporary increase in the fed funds rate on inflation and the unemployment rate? This paper examines the relationship between the fed funds rate, inflation, and unemployment rate over the 1970-2014 period. By using a quarterly data set, this study posits answers to these questions.

The FOMC sets the pace of U.S. monetary policy and provides a target for the fed funds rate. Two primary factors (among several others) that influence FOMC decisions are the inflation rate and the unemployment rate. The FOMC has even stated that the methods of monetary policy may be altered it' the inflation rate runs persistently above or below the 2 percent level. Furthermore, one of the FOMC's longterm goals is to help the economy return to the full employment level. If the economy persistently suffers a high unemployment rate, this may influence future FOMC decisions. A famous rule, or what is now commonly referred to as the Taylor rule, has become a popular way to conceptualize monetary policy decision making and evaluate the appropriate tenor of monetary policy against a consistent set of benchmarks [Bernanke 2010]. The Taylor rule suggests that the fed funds rate depends on inflation expectations and the output gap [see Taylor 1993 for more detail]. Stock and Watson [2001] utilized a modified Taylor rule, which utilizes the fed funds, inflation, and unemployment rates to link the relationship between the three variables. We follow the Stock and Watson approach and utilize a modified version of the Taylor rule to evaluate the effect of a change in the red fund rate on inflation and the unemployment rate.

To test a statistical relationship between interest rates, inflation, and the unemployment rate, two econometric techniques are employed. First, we apply a Granger causality test [Granger 1969] to determine a causal relationship between the variables. The major reason to test the causal relationship is that the FOMC changes its target for the fed funds rate to bring inflation and the unemployment rates closer to mandate-consistent levels. The Granger causality test would determine whether there is statistical evidence that the fed funds rate is causing, movements in inflation and the unemployment rate. The second econometric technique, Vector Autoregrcssions (VARs) modeling proposed by Sims [1980], estimates the likely effect of a 1 percentage-point increase in the fed funds rate on the rate of inflation and the unemployment rate.

Both the VAR and Granger causality test are estimated using the structural VAR approach. Typically, a structural VAR model is built based on an economic theory. That is, we include real GDP growth (year-over-year percent change, or YoY) in the model to capture aggregate demand and the growth in the S&P 500 index (YoY) to incorporate financial sector activity. Housing starts (YoY) are utilized in the model to represent housing sector activity. As a structural VAR model, such as our model, includes information from major sectors of an economy, it is therefore likely to provide reliable statistical results [Stock and Watson 2001].

Some argue that the U.S. economy, in particular inflation and the labor market, performed differently in the post-1990s era compared with the pre-1990s era [Bernanke 2004]. First, the operation of this market was altered by the passage of the Financial Institutions Reform and Recovery Act of 1989. The focus of this Act was to restructure the operations of the savings and loan industry. Effectively, the Act also allowed for a more efficient allocation of credit and deposits in the financial system. In this sense, deposit gathering more closely reflected the need for capital and the pursuit of federal funds for short-term financing. Second, beginning in the mid-1980s, the conduct of monetary policy had drifted away from the targeting of nonborrowed reserves and the money supply toward the use of the federal funds rate. (1) Our choice of basing our work on the period since 1990 reflects the implementation of the 1989 Act and the shift of Federal Reserve policy to utilizing the fed funds rate to target specific economic outcomes.

For instance, the mean and standard deviation of the inflation rate for the 1990:Q1-2014:Q1 period are 2.12 and 0.96 percent, respectively, which are smaller than the average and standard deviation for the complete period (3.73 and 2.61 percent, respectively). This suggests that the post-1990s era experienced a lower inflation rate, on average, and the volatility around the average inflation was also lower than in the complete sample period. The post-1990s era includes what is often known as the Great Moderation because the volatility in the major U.S. macroeconomic variables declined [Bernanke 2004]. The labor market also showed different characteristics in the post-1990s era compared with prior history. The past three recoveries are considered "jobless" recoveries by some observers, suggesting a possible structural break in the labor market behavior since the 1990s. Therefore, it would be important to test the idea of a structural break because for policy recommendations an estimated relationship should be consistent among all subsamples in order to base any decisions on that relationship. The above scenario may also reflect the Lucas Critique [Lucas 1976], which suggests that if the estimated relationship changes whenever policy changes (or sample period changes), then policy conclusions based on the estimation are misleading.

To test the idea of a structural break, we run the Granger causality test and VAR modeling using the data set for several subsample periods, that is, the post-1990s era, 1990-2007:Q3, and 2007:Q4-2014:Q1 periods. The idea behind using the subsamples is that if the results are statistically different from the complete period, then this difference suggests that the relationship between the three variables has changed since 1990.

Our statistical analysis found mixed evidence about the relationship between interest rates, inflation, and the unemployment rate. That is, during the complete sample period (1971-2014), the fed funds rate was found to Granger-cause the inflation rate and unemployment rate. However, when narrowing the time period to 1990-2014 and 1990-2007, the fed funds rate does not produce the same Granger causality on inflation. Yet, for the same periods, the fed funds rate does Granger-cause the unemployment rate.

In the period since the Great Recession (2007: Q4-2014:Q1), the fed funds rate does not Granger-cause inflation or the unemployment rate. This suggests that a change in the fed funds rate may not be the key driver of inflation rates movements and, therefore, the link between policy changes in the fed funds rate and the target inflation rates is not as straightforward as previously believed.

Moreover, our statistical analysis indicates that the inflation and unemployment rates Granger-cause movements in the fed funds rate for the 1971-2014 period. As the fed funds rate is discretionary, this suggests a reaction by the FOMC. However, similar results could not be found for subsamples. For the 1990-2014 period, only inflation (and not the unemployment rate) Granger-causes the fed funds rate, and for the 1990-2007:Q3 and 2007:Q4-2014:Q1 periods, neither inflation nor unemployment rate Granger-causes fed funds rate. This suggests that these variables are not statistically useful to predict actions of the FOMC. However, it is also possible that causality exists but is less direct than can be captured by an econometric model.

We also found mixed evidence about the effect of a 1 percentage-point increase in the fed funds rate on inflation and the unemployment rate. When looking at the 1971-2014 period, an increase in the interest rate is associated with a rising inflation rate, except for the first quarter when it decreases. The interest rate hike is also associated with a declining unemployment rate. For the 1990-2014 and 1990-2007 periods, an increase in the fed funds rate leads to a decline in the inflation rate and rise in the unemployment rate, on average.

Both the Granger causality test and VAR analysis suggest that the relationship between interest rates, inflation, and the unemployment rate is different in the post-1990s era compared with the complete sample period. Put differently, would a hike in the fed funds rate increase inflation pressure (as suggested by the complete sample period analysis) or would it reduce inflation pressure (as suggested by the post-1990s subsample analysis)? This finding is vital for decision makers. For policy recommendations, an estimated relationship should be consistent among all subsamples in order to base any decisions on that relationship.

The rest of the paper is organized as follows: Section 2 provides the theoretical review of the relationship between interest rates, inflation, and unemployment rates. Section 3 reviews the econometric setup of the paper. Section 4 explains the issues related to the data. The results are discussed in Section 5. The concluding remarks are gathered in Section 6.

2. The Theoretical Relationship between Interest Rates, Inflation, and Unemployment Rates

The primary goals of the FOMC are to foster maximum employment and price stability. One of the key tools of the FOMC to meet its goals is the fed funds rate, which affects other short-term interest rates, long-term rates, exchanges rates, and several economic variables, including employment, output, and inflation rates.

As of 2012, the FOMC has announced a specific long-run target for the inflation rate of 2 percent, measured by the Personal Consumption Expenditures (PCE) deflator. The FOMC has even stated that the methods of monetary policy may be altered beyond manipulation of the fed funds rate if the inflation rate runs persistently above or below 2 percent. Between December 2012 and March 2014, the FOMC also included a threshold for the unemployment rate, where the fed funds rate would remain on hold as long as unemployment was above 6.5 percent. If the economy persistently suffers a high unemployment rate, this scenario may also influence future FOMC decisions. Typically, during a recession, as the unemployment rate moves upward, the FOMC reduces the fed funds target rate to stimulate the economy and bring the unemployment rate down to the preferred natural level (2).

On the one hand, the FOMC adjusts the fed funds target to influence inflation and the unemployment rate. However, actual inflation and the actual unemployment rate are two key determinants of the FOMC's decisions to change the fed funds rate. For instance, Taylor [19931 suggested that deviations of inflation and unemployment from some desired level would lead to changes in the fed funds rate, in what is now commonly referred to as the Taylor rule. Mankiw [2001] concluded that inflation and the unemployment rate are important factors of fed funds rates in what is sometimes known as the Mankiw rule. Stock and Watson [2001] developed a modified Taylor rule, which utilized inflation and the unemployment rate as determining factors of the fed funds rate in order to estimate an econometric relationship between the three variables.

Although the fed funds target rate, inflation, and the unemployment rate are *related theoretically, it is important to quantify the precise statistical relationship between the three variables. Specifically, do the data support a causal relationship between the variables? That is, do changes in the fed funds rate affect inflation and the unemployment rate and vice versa? In addition, how much do inflation and the unemployment rate change in response to a temporary 1 percentage-point increase in the fed funds rate?

Over time, many economic and financial variables change their behavior as some become more (or less) volatile compared with the past. This changing nature of variables may impact the theoretical relationship between them. For example, the volatility in the major U.S. macroeconomic series such as real GDP, industrial production, and inflation declined in the post-1990s era--the Great Moderation [Bemanke 2004]. For instance, the mean and standard deviation of the inflation rate, measured by the PCE deflator, for the 1990-2014 period are 2.12 and 0.96 percent, respectively, which are smaller than the average and standard deviation for the complete period (1971-2014), which are 3.73 and 2.61 percent, respectively. This suggests that the post-1990s era experienced a lower inflation rate, on average, and the volatility around the average rate of inflation was also lower than that of the complete sample period.

The labor market also showed different characteristics in the post-1990s era compared with the past. The last three recoveries are considered "jobless" recoveries by some observers, suggesting a possible structural break in the labor market behavior since the 1990s. Silvia [20061 provided a detailed discussion about the structural changes in the U.S. labor market in a global context even before the shock of the Great Recession.

Summing up, theoretical literature suggests a relationship between the fed funds target rate, inflation, and the unemployment rate. In addition, the traditional relationship between these variables may have changed since the 1990s. Therefore, it is vital for decision making to quantify (1) whether the fed funds rate is useful to predict movements in the inflation and the unemployment rate, (2) whether inflation and the unemployment rate statistically cause changes in the fed funds rate (possibility of a two-way causality), and (3) whether the relationship has changed since the 1990s. The next section discusses econometric techniques that are utilized to answers these questions.

3. Econometric Setup

The Granger causality test

First, we test whether the fed funds rate is statistically useful in explaining movements in the inflation and the unemployment rate. The Granger causality test [Granger 19691 is utilized to determine the direction of a relationship between variables. Granger causality is a statistical concept of causality that is based on prediction. According to Granger causality, if a variable X, "Granger-causes" a variable Y,, then past values of X, should contain information that helps predict Y, above and beyond the information contained in past values of Y, alone [Granger 1969]. The Granger causality test also indicates the direction of the causality, that is, whether it is a one-way or two-way causality. For instance, if X, "Granger-causes" Y, but Y, does not "Granger-cause" X, then the relationship would be called one-way causality. On the other hand, if X, "Granger-causes" Y, and Y, also "Granger-causes" X, then it indicates two-way causality. For instance, in the present case, we test whether the fed funds rate Granger-causes inflation and the unemployment rate. That is, whether the fed funds rate helps to increase predictability of the inflation and the unemployment rate. We can also test whether there is one-way causality (from the fed funds rates to inflation and the unemployment rate only) or two-way causality (inflation and the unemployment rate also Granger-cause fed funds rates). (3)

The VAR and the impulse response function

What would be the likely effect of a 1 percentage-point increase in the fed funds rate on inflation and the unemployment rate? To answer this question, we turn to the VAR modeling methodology. The beauty of VARs is that they are simple statistical representations of economic systems as they rely only on the variables that comprise the system and a few lagged values of those variables. In addition, VARs can be "shocked" to show how all the variables in the model respond to a change in one of the other variables. The way the variables respond over time to a change in the "shocked" variable is called an impulse response function (IRF).

Sims [1980] introduced the VAR modeling approach as an alternative to the large-scale structural, macroeconometric model. The basic idea behind a VAR approach is that instead of including hundreds of variables in a model, we can include a handful of variables (sometimes eight variables) to represent major sectors of an economy and then that model can be utilized for forecasting and policy analysis.

A traditional VAR of n variables will consist of n equations, one equation for each variable. Each equation includes a constant and lag(s) of n variables, including lag(s) of the left-hand-side variable. The lag order, how many lags of a variable, is denoted by "P". Therefore, a VAR (P) of n variables indicates that up to p lags of each variable are utilized in each equation.

A simple example of a two-variable VAR model that includes one lag of each variable is shown below. We will call this example VAR(1), because P, the lag period, equals 1.

[Y.sub.t] = [[alpha].sub.0] + [[alpha].sub.1][y.sub.(t-1)] - + [[alpha].sub.2][X.sub.(t-1)] + [[epsilon].sub.1t]

[X.sub.t] = [[beta].sub.0] + [[beta].sub.1][Y.sub.(t-1)] + [[beta].sub.2][X.sub.(t-1)] + [[epsilon].sub.2t]

In an n-variables VAR model, an IRF shows the effect of a change in the lagged value of one variable in the current period on all variables in the model in the next period, assuming that the change will disappear in the subsequent periods. Furthermore, we keep the VAR errors for other (n-1) variables equal to zero for the current period; that is, actual values are equal to estimated values. That would allow generating the effect of the change in one variable on all other variables.

Testing causality and IRF using a structural VAR approach

A structural VAR approach is utilized to test Granger causality and to estimate IRFs for the three variables. A structural VAR model includes variables from major sectors of an economy in addition to the variables of interest.4 For instance, our objective is to test a statistical relationship between the fed funds rate, inflation, and the unemployment rate, but, in addition to these variables, we also include growth in real GDP, the S&P 500 index, and housing starts (measured by the year-over-year percent change) in the model. The major reason of using a structural VAR approach to test Granger causality and to estimate IRF is that it improves model fit and provides useful and reliable results [Greene 2011].

Sometimes. two or more variables are influenced by another variable, known as a third factor, which suggests movements in the third factor may statistically associate with movements in the variables of interest. Excluding the third factor from the model would cause omitted variable bias in the parameter estimates.5 For example, during recessions, the unemployment rate tends to rise whereas the fed funds target rate and inflation tend to decline. That is, all three variables are influenced by the overall economy (or aggregate demand), which is one reason we include real GDP growth in the model (see the next section for more detail about other variables). Therefore, a structural VAR approach, depending on how well it is structured, would produce reliable statistical results to examine whether there is a relationship between the fed funds rate, inflation, and the unemployment rate.

4. The Data and Implementation Strategy

This study uses a quarterly data set for 1971:Q1-2014: Qi. The fed funds target rate is a key determinant of short-term interest rates (and to some extent long-term rates as well), and it is the overnight lending rate at which depository institutions lend each other reserves. The PCE deflator, year-over-year percent change (YoY), is the Federal Reserve's preferred measure of inflation, which is why we utilize it instead of other inflation measures, such as the Consumer Price Index or Producer Price Index. The final variable of interest is the unemployment rate, measured as the total number employed as a percentage of the civilian labor force. The real GDP percentage change (YoY) is a good proxy for aggregate demand. The financial market is an important element of the U.S. economy, and therefore we include percentage change in the S&P 500 index (YoY) in the model to capture financial sector activity. The housing sector played a vital role in the Great Recession and we also include the percentage change in housing starts (YoY) in the model to incorporate housing sector activities. Therefore, the final VAR model includes the above-mentioned six variables.

As mentioned earlier, due to the changing nature of the variables, the relationship between these variables may have changed over time. In the case of the U.S. economy, for example, the volatility in the major macroeconomic series such as real GDP, industrial production, and the inflation rate declined in the post-1990s era. Since December 2008, the fed funds target rate has remained at 0-25 basis points, the lowest level in history. In addition, the inflation rate and the labor market performed differently in the post-1990s era compared with the pre-1990s period. Therefore, it is important to determine, statistically, the possibility of a structural break in the relationship between interest rates, inflation, and the unemployment rate.

The Granger causality test and VAR approach are employed to determine a statistical relationship. Four different time periods are covered in this study. The first period includes data for 1971:Q1-2014:Q1 and it is termed the complete sample. The second period is a subsample for the 1990:Q1-2014:Q1 span. The reason to rerun the econometric analysis using this period is to determine whether the statistical results based on the subsample are different compared with the complete period (1971-2014). If so, the result would suggest the possibility of a structural break. The 19902014 period contains the Great Moderation and the Great Recession; and since the Great Recession, many macroeconomic and financial variables performed differently from the past. Therefore, the third subsample utilizes the time period 1990:Q1-2007:Q3, before the Great Recession. The final subsample is for the time period since the Great Recession, 2007:Q4-2014:Q1. The data set is obtained from the LHS Global Insights.

5. The Results

Testing the direction of the relationship: Cause and effect discussion

First, using the structural VAR approach, we test whether the fed funds rate is statistically useful in explaining movements in the rate of inflation and the unemployment rate. That is, whether the fed funds rate helps to increase predictability of the inflation and the unemployment rate. The Granger causality test is utilized to identify the direction of a relationship between variables.

Table 1 provides the Granger causality results to determine the relationship between the fed funds rate, inflation, and the unemployment rate over the entire sample period (1971-2014). As shown, the fed funds rate Granger-causes both the inflation and the unemployment rate. Therefore, the fed funds rate can be useful in predicting the rate of inflation and the unemployment rate. Inflation and the unemployment rate also Granger-cause the fed funds rate. That means inflation and the unemployment rate are useful predictors of the direction of the fed funds rate, which is to be expected as the federal funds rate is the means by which the FOMC reacts to changes in these variables.

Table 1. The Granger Causality Test: Regressor 1971:Q1 -2014:Q1

Regressor                          Dependent Variable

                   Fend funds rate  PCE deflator  Unemployment rate

Fed funds rate            NA          0.10***             0.00*
PCE defliator          0.00*               NA             0.00*
Unemployment rate     0.02**            0.01*                NA
rate

Note: real GDP (YoY), S&P 500 index (YoY), and housing starts
(YoY) are also included in the model.

* Significant at 1 percent, ** Significant at 5 percent

***Signiigicant at 10 percent.


Table 2. The Granger Causality Test: 1990:Q1 -2014:Q1

Regressor                               Dependent Variable

                   Fed funds rate     PCE deflator  Unemployment rate

Fed funds rate           NA                0.3           0.00*
PCE deflator         0.04**                 NA           0.00*
Unemployment rate      0.46             0.04**              NA
rate

Note: -real GDP (YoY), S&P 500 index (YoY), and housing starts
(YoY) are also included in the model.

* Significant at 1 percent, ** Significant at 5-.percent,
***Significant at 10 percent.


Next, to test the idea of a structural break, we run Granger causality tests with three subsamples for the post-1990s era. The idea behind using these subsamples is that if the results are statistically different from the complete period, then this difference suggests that the relationship between the three variables has changed since 1990. The Granger causality results based on the 1990-2014 period are reported in Table 2 and indicate that the fed funds rate helps in the predictability of the unemployment rate, but not the inflation rate. On the other hand, inflation rates provide a good prediction for the fed funds rate, while the unemployment rate does not Granger-cause fed funds in the post-1990s era.

The third subsample utilizes the time period before the Great Recession (1990-2007:Q3). Results are reported in Table 3. The fed funds rate Granger-causes the unemployment rate but does not Granger-cause inflation rates. Furthermore, inflation and the unemployment rate do not Granger-cause the fed funds for this period. The final Granger causality test analyzes the experience since the Great Recession (2007:Q4-2014:Q1) period and results are reported in Table 4. There is no causality between the fed funds rate, inflation, and the unemployment rate in this period. However, the time span for the final data set is very short and Granger causality results are sensitive to the shorter period. Another reason for the apparent lack of a relationship is that the fed funds target rate has been in the 0-0.25 percent range since December 2008, and with basically no significant changes in the rate, there may not appear to be any causality findings.

Table 3. The Granger Causality Test:1990:Q1 -2007:Q3

Regressor           Dependent Variable

              Fed funds     PCE    Unemployment
                 rate    deflator      rate

Fed funds rate    NA        0.51       0.00*
PCE deflator     0.29         NA       0.00*
Unemployment     0.34       0.49          NA
rate

Note: real GDP (YoY), S&P 500 index (YoY), and housing
starts (YoY) are also included in the model.

* Significant at 1 percent, ** Significant at 5 percent,
***Significant at 10 percent.


Table 4. The Granger Causality Test: 2007:Q4-2014:Q1

Regressor             Dependent Variable

                 Fed funds      PCE     Unemployment
                   rate      deflator      rate
Fed funds rate      NA         0.31        0.68
PCE deflator      0.72           NA     0.06***
Unemployment      0.87          0.2          NA
rate

Note: real GDP (YoY). S&P500 index (YOY), and housing
starts (YoY) are also included in the model.

* Significant at 1 percent, ** Significant at 5 percent.
***Significant at 10 percent.


Meanwhile, our statistical analysis indicates that the rate of inflation and the unemployment rate Granger-cause movements in the fed funds rate for the 1971-2014 period. However, similar results could not be found for subsamples. That is, for the 1990-2014 period, only inflation (and not the unemployment rate) Granger-causes the fed funds rate, and for the 1990: Q1-2007:Q3 and 2007:Q4-2014:Q1 periods neither inflation nor unemployment Granger-causes the fed funds rate. This suggests that these variables are not statistically useful to predict movements in the fed funds rate, which, after all, is a matter of discretion on the part of the FOMC. However, it is also possible that causality exists but is less direct than can be captured by an econometric model.

In sum, the Granger causality results suggest that since the 1990s the fed funds rate may influence the unemployment rate and not inflation as they have in the past and certainly not as much as commonly perceived in the financial markets. One possible reason for the no-causality findings between the fed funds and inflation rate is that since the 1990s, on average, inflation rates have stayed close to the FOMC's target of 2 percent. It is possible that causality exists but is less direct than can be captured by using only the lagged values of these variables.

The fed funds rate, inflation, and the unemployment rate: The structural VAR and IRF

The structural VAR approach is utilized to estimate the IRF. Using the complete sample period (1971-2014), we "shocked" (increased) the fed funds rate by 1 percentage point to examine the effect on inflation and the unemployment rate. Furthermore, the total effect of a change in the fed funds rate on the variables may be distributed over a prolonged period of time. Therefore, we generate the effect of a change in the interest rates in the current quarter on the rate of inflation and the unemployment rate over the next 12 quarters. Figure 1 shows what effect a temporary 1 percentage-point increase in the fed funds rate has on inflation and the unemployment rate. The hike in the interest rate reduced the inflation rate only for the first quarter (reduced by 0.02 percentage points) and after that inflation increased 0.1 percentage points in the second quarter and continued on an increasing trend, contrary to expectations. The largest increase in the inflation rate was seen in the 12th quarter with a jump of 0.8 (rounded upward) percentage points. The upward shock in the fed funds rate reduces the unemployment rate by 0.4 (rounded downward) percentage points--again contrary to expectations--in the first quarter, with the largest drop of 0.8 (rounded upward) percentage points experienced in the sixth quarter.

Typically, during an economic expansion, the FOMC tends to raise the fed funds target rate to combat inflation pressure. At the same time, as the economy is growing, there may be downward pressure on the unemployment rate. The IRF results are consistent with a declining unemployment rate (but not with the causality expected of an increase in the fed funds rate), but inflation has an increasing trend, possibly suggesting that in our sample period traditional tools of monetary policy were unable to reduce inflation pressures and that any interest rate effect is overwhelmed by other relationships in the structural model.

The IRFs are also calculated for the three subsamples (1990-2014, 1990-2007:Q3 and 2007:Q4-2014) of the post-1990s era. There are some noticeable differences in Figure 2 for the 1990-2014 period compared with Figure 1 for the entire period. The post-1990s era shows that tightening of monetary policy is attached to softening inflation pressure. That is, a hike in the fed funds rate reduces the inflation rate by 0.2 percentage points in the first quarter, while the largest decline is reported at 0.6 percentage points in the 10th quarter. The figure shows a rising trend in the unemployment rate, 0.1 percentage points for the first quarter and the largest increase at 0.4 percentage points in the 12th quarter. Figure 2 shows a rise in the unemployment rate, instead of a drop as shown in Figure 1.

Figure 3 is based on the 1990:Q1-2007:Q3 period, which excludes the post-Great Recession era. The inflation rate rises for the first three quarters followed by a declining trend. The unemployment rate, on the other hand, drops for the first quarter and then starts rising for the next 11 quarters. Some noticeable observations in Figure 3 compared with Figures 1 and 2 are, first, magnitudes of changes (in absolute terms) in the inflation rate are much smaller (largest change is 0.3) in Figure 3 compared with Figure 1(largest change is 0.8) and Figure 2 (largest change is 0.6). Second, Figure 3 shows the largest increase of unemployment by 0.9 percentage points compared with the largest drop of 0.8 in Figure 1 and the largest increase of 0.4 percentage points in Figure 2.

The IRFs based on the post-Great Recession era (2007:Q4-2014:Q1) are shown in Figure 4. The inflation rate shows a mixed response and the magnitude of changes are very small, with the largest drop of 0.1 percentage points in the third quarter. The unemployment rate also shows a mixed response and the magnitude of changes is very small, with the largest drop of 0.1 percentage points in the first quarter. Given the short history of the data set for Figure 4, results may be sensitive as a VAR analysis would provide more reliable results using a longer span of time compared with a short history.

An important factor may be that the fed funds target rate since 2008 has been at the lowest level (in the 0-0.25 percent range) since data were first recorded in 1971 [Federal Reserve Bank of New York Current] and has seen the longest period without a target rate change (December 2008 to present).

In sum, the IRF results suggest that in the post-1990s era, a hike in the fed funds rate implies, on average, declines in inflation pressure and a rising unemployment rate. These results are opposite of those results that are based on the complete sample period of 1971-2014.

Breaking down history: The Lucas critique and policy recommendations

Both the Granger causality test and VAR analysis suggest that the relationship between interest rates, inflation, and the unemployment rate has been different in the post-1990s era compared with the complete sample period. This finding is vital for decision makers. For policy recommendations, an estimated relationship should be consistent among all subsamples in order to base any decisions on that relationship. Put differently, in the present case, would a hike in the fed funds rate increase inflation pressure (as suggested by the complete sample period analysis) or would it reduce inflation pressure (findings of the post-1990s subsamples)?

The above scenario may also reflect the Lucas Critique [Lucas 1976]. In simple terms, the Lucas Critique suggests that if the estimated relationship (or parameters) changes whenever policy changes (or sample period changes), then policy conclusions based on the estimation are misleading. In sum, the relationship between interest rates, inflation, and the unemployment rate is sensitive to the policy regime. This may suggest that the effect of the fed funds target rate on inflation and the unemployment rate may not be as straightforward as suggested by the conventional economic theory.

6. Conclusion

One of the important tools of monetary policy to meet the FOMC's long-term goals is the fed funds target rate. In theory, all three variables are related. However, our statistical analysis suggests that the relationship between the fed funds rate, inflation, and the unemployment rate is different in the post-1990s era compared with the complete sample period. This finding is vital for decision makers. For policy recommendations, an estimated relationship should be consistent during subsamples and if the relationship is not consistent between different periods then we must be careful to make policy recommendations based on that relationship.

These findings are possible reasons why the Federal Reserve Board has employed different tools of monetary policy during the past seven years. These tools include both the traditional, such as the fed funds target rate and the nontraditional, such as large-scale asset purchases and forward guidance.

REFERENCES

Bemanke, Ben. 2004. The Great Moderation, February 20, 2004. Remarks at the Meetings of the Eastern Economic

Association, Washington, DC, http://www.federalreserve.gov/Boarddocs/Speeches/2004/20040220/.

Bernanke, Ben S. 2010. Monetary Policy and the Housing Bubble. The Annual Meeting of the American Economic

Association, Atlanta, Georgia, January 2010, http://wwwlederalreserve.govinewsevents/speech/bemanke20100103a .htm.

Board of Governors of the Federal Reserve System, 2005. The Federal Reserve System: Purposes & Functions, http:// www.federaireserve.gov/pf/pdf/pf_complete.pdf.

Federal Reserve Bank of New York, Current, Historical Changes of the Target Federal Funds and Discount Rates, http://www.newyorkfed.orgimarkets/statistics/dlyrates/fedrate.html.

Granger, C.W.J. 1969. "Investigating Causal Relationships by Econometric Models and Cross-Spectral Methods." Econometrica, 37(3):-424 438.

Greene, William H. 2011. Econometric Analysis, 7th ed. Prentice Hall.

Lucas, Robert. 1976. "Econometric Policy Evaluation: A Critique."Carnegie-Rochester Conference Series on Public Policy, 1( 1 ): 19-46.

Mankiw, Gregory N. 2001. U.S. Monetary Policy during the 1990s. NBER Working Paper No. 8471.

Silvia, John. 2006. "Domestic Implications of a Global Labor Market." Business Economics. 41(3): 23-29.

Silvia, John E., Azhar Iqbal, Sam Bullard, Sarah Watt and Kaylyn Swankoski. 2014. Economic and Business Forecasting: Analyzing and Interpreting Econometric Results. Wiley.

Sims, Christopher. 1980. "Macroeconomics and Reality." Econometrica, 48(1): 1-48.

Stock, James and Mark Watson. 2001. "Vector Auto-regressions." Journal of Economic Perspectives, 15(4): 101-115..

Taylor, John. B. 1993. "Discretion versus. Policy Rules in Practice." Carnegie-Rochester conference Series on Public Policy, 39: 195-214.

Caption:Figure 1. A Shock to Fed Funds Target Rate (1971-2014)

Caption:Figure 2. A Shock to Fed Funds Target Rate (1990-2014)

Caption:Figure 3. A Shock to Fed Funds Target Rate (1990-2007:Q3)

Caption:Figure 4. A Shock to Fed Funds Target Rate (2007:Q4-2014)

<FOOTNOTE>

(1.) iuBeginning in the mid-1980s, spreading doubts about the financial health of some depository institutions led to an increasing reluctance on the part of many institutions to borrow at the discount window, thus weakening the link between borrowing and the federal funds rate. Consequently, the Federal Reserve increasingly sought to attain a specific level of the federal funds rate rather than a targeted amount of borrowed reserves. In July 1995. the FOMC began to announce its target for the federal funds rate-[Board of Governors 2005, p. 29, www .federalreserve.gov/pf/pdf/pf complete.pdfj

(2.) The Congressional Budget Office provides an estimate of the natural unemployment rate that is currently at 5.5 percent.

(3.) The statistical techniques here are covered in more detail in econometrics text books and in Silvia and others [20141.

(4.) See Stock and Watson [2001] for more detail about a structural VAR approach.

(5.) For more detail about omitted variable bias and reliable results, see Chapter 8 of Greene [2011].

*John Silvia is a managing director and the chief economist for Wells Fargo, continuing the position he has held since he joined Wachovia in 2002 as the company's chief economist. Prior to his current position, he worked as senior economist for the U.S. Senate Joint Economic Committee and chief economist for the .U.S. Senate Banking, Housing and Urban Affairs Committee. Before that, he was chief economist of Kemper Funds and managing director of Scudder Kemper Investments, Inc. Silvia is currently the President of the National Association of Business Economics (NABE) and former president of the Charlotte Economics Club. He holds B.A. and Ph.D. degrees in economics from Northeastern University and a Master's degree in economics from Brown University.

Azhar Iqbal is an econometriciarian at Wells Fargo, responsible for providing quantitative analysis to the Economics group and modeling and forecasting macro and financial variables. Before joining Wells Fargo in 2007, he was an economist and course instructor at the Applied Economics Research Center at the University of Karachi in Pakistan. He has also worked as an economist at the United Nations, Arif-Habib Investment Bank, and for Government of Pakistan-funded projects. Iqbal received his Bachelor's degree in economics from the University of Punjab and has Master's degrees from the State University of New York at Albany, University of Karachi, and the University of the Punjab.
COPYRIGHT 2014 Springer
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:federal
Comment:Is the fed funds rate still effective?(federal )
Author:Silvia, John and Iqbal, Azhar
Publication:Business Economics
Geographic Code:1USA
Date:Oct 1, 2014
Words:6795
Previous Article:Is Fiscal Stimulus a good idea?
Next Article:Confronting the gray market problem.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |