Printer Friendly

Federal Reserve transcript publication and regional representation.

I. INTRODUCTION

This article looks at disagreement within the Federal Reserve's monetary policy committee, the Federal Open Market Committee or FOMC, following a change in transparency practices taken in 1993 to publish verbatim transcripts of FOMC meetings. The article examines whether disagreement during FOMC deliberations before and after the change in transparency practices provides evidence of regional heterogeneity.

The structure of the FOMC is well known: the 12-member committee is composed of the seven board members on the Board of Governors in Washington DC and five of the presidents from the 12 Federal Reserve Banks (for additional detail, see Purposes and Functions, 2005). While the seven board members and the president of the Federal Reserve Bank of New York vote at every FOMC meeting, the other four voting presidents are determined according to a legislated annual rotation. (1) All 19 officials, whether voting or not, participate in the policy deliberations at each meeting.

Each FOMC member has some, at least nominal, affiliation with one of the 12 Federal Reserve regions that were established at the time the Fed was founded. The bank presidents are appointed to their positions by the bank's board of directors (made up of representatives from local commercial banks and the public). Each board member also has some affiliation with one of the districts, and the law stipulates that no two board members can come from the same district.

A number of authors have looked at a possible regional dimension to the FOMC's voting patterns. The early literature on this issue was concerned mainly with the differences between the votes of board members and bank presidents and ignored the regional affiliations of the former (Belden (1989), Havrilesky and Schweitzer (1990), and Gildea (1990, 1992) are examples of this work). More recent literature by Meade and Sheets (2005b) and Chappell, McGregor, and Vermilyea (2008) has provided statistical evidence that regional economic information (as measured by a regional unemployment rate) affects the votes cast by FOMC officials and the policy preferences they voice during meeting deliberations. These recent studies tell us that regional effects are important not only for bank presidents but for board members as well.

In this article, I look at the decisions of FOMC members to see whether these decisions were influenced by regional effects before and after the major change in transparency practices that occurred in 1993. Before 1993, the FOMC published individual votes of committee members as well as summary minutes of meetings, but it did not publish full transcripts that would allow outside observers to determine exactly what individual committee members said about their policy preferences. Under pressure from the U.S. Congress during the fall of 1993 to increase the transparency of the monetary policy process, the FOMC agreed to release lightly edited transcripts of each meeting after a 5-yr delay. While many officials knew at the time they agreed to release the transcripts that FOMC meetings were tape-recorded, most thought that the recordings were used to prepare meeting minutes and then destroyed. Thus, because transcripts exist from a time when meeting participants did not know that their deliberations would be made public, it is possible to compare the meeting discussions of policymakers before 1993, when they believed that their remarks were private, and after 1993, when they knew that all statements would eventually be made public. Currently, complete transcripts are available for nearly all FOMC meetings from 1978 through 2003.

Geraats (2002) identifies two possible effects of increased monetary policy transparency: a reduction or elimination of information asymmetries between the central bank and the public and an incentive for policymakers to behave strategically. The incentive effect can arise when policymakers have both a public and a private objective--for instance, a desire to make the "correct" policy decision as well as obtain reappointment or enhance future job prospects (Ottaviani and Sorensen, 2001, model this incentive effect for a committee that deliberates in public). While an increase in transparency can raise welfare by reducing the information asymmetry, strategic behavior could potentially offset the welfare gain if policymakers withhold information during their deliberations in order to enhance their reputations. Gersbach and Hahn (2004, 2005, 2008) model the effects of publishing monetary policy votes and show that the net effect of increased transparency depends on whether the public objective of the voter is sufficiently strong relative to his or her private objective. Meade and Stasavage (2008) model the effects of opening the FOMC's deliberations to public view and provide empirical evidence to support the hypothesis that the publication of transcripts made policymakers less willing to voice disagreement with the chairman's policy proposal. This article adds to that work by examining whether regional variables are important to the analysis and whether the transcription effects are robust to the inclusion of regional indicators. The estimation results indicate that transcription effects are indeed robust, regardless of the regional indicator used, and the size of Federal Reserve districts matters during FOMC deliberations in that larger districts may be more likely to voice agreement with a given policy proposal. The next section reviews the size of the 12 Federal Reserve regions and their voting patterns. Section III discusses the data, sets up the estimation equation, and describes the empirical results. Section IV concludes.

II. REGIONAL REPRESENTATION ON THE FOMC

Table 1 examines how FOMC votes have lined up with Federal Reserve regions over a long period (1968-2004) taking into account the regional affiliations of both board members and bank presidents. The New York district has had by far the largest voting share, with an average 1.93 votes at each FOMC meeting, for a total of 733 votes. The New York Fed president or his representative cast a vote at all the meetings, and a board member affiliated with the New York region voted 93% of the time. Furthermore, this voting share probably understates the influence of the New York district in the formulation of monetary policy because three of the Fed chairmen during this period (William Martin, Arthur Burns, and Alan Greenspan) were appointed from the New York district. After New York, the regions can be separated into three groups: Chicago and Richmond have each cast about 1 1/4; votes at FOMC meetings; Boston, Kansas City, Dallas, Philadelphia, and San Francisco have cast roughly one vote per meeting; St. Louis, Atlanta, Cleveland, and Minneapolis have had the least voting power, casting only about 1/2; vote per meeting.
TABLE 1

Average Votes Per FOMC Meeting, 1968-2004 (a)

District       Total  Board  Bank  Total Votes

New York       1.93   0.93   1.00      733
Chicago        1.27   0.79   0.48      485
Richmond       1.20   0.89   0.31      455
Boston         1.05   0.68   0.37      398
Kansas City    1.00   0.63   0.37      380
Dallas         0.98   0.65   0.33      371
Philadelphia   0.95   0.63   0.32      362
San Francisco  0.85   0.54   0.31      324
St. Louis      0.64   0.31   0.33      244
Atlanta        0.60   0.26   0.34      228
Cleveland      0.51   0.00   0.51      195
Minneapolis    0.44   0.12   0.32      170

(a) Calculated for 380 FOMC meetings, including scheduled meetings and
conference calls.

Source: Meade and Sheets (2005a).


The data demonstrate that the Fed regions are unequal in terms of voting shares, which is to be expected since the regions themselves differ in size. Table 2 looks at region size in terms of several measures. Total assets on the balance sheet of each Federal Reserve Bank provides an imperfect but rough gauge of the financial size of each region. In 1990, New York was by far the largest region (about one-third of total Reserve Bank assets), followed by Chicago (13% of the total). Six of the 12 regions--Atlanta, Dallas, Kansas City, Minneapolis, Philadelphia, and St. Louis--were quite small, each accounting for no more than 5% of the total.
TABLE 2

Rankings of Federal Reserve Districts

                         Measure of Size                     Note

               Assets  Population  Real GDP  Votes Per     GDP Per
                 (a)     (1990)     (1990)     Meeting    Capita (a)
               (1990)                        (1968-2004)  (Millions)

Boston            5        10          8           4        26,943
New York          1         4          2           1        29,797
Philadelphia     10        11         10           7        24,313
Cleveland         6         7          7          11        21,863
Richmond          4         5          5           3        24,876
Atlanta           7         2          4          10        20,994
Chicago           2         3          3           2        22,706
St. Louis        11         9         11           9        20,272
Minneapolis      12        12         12          12        22,066
Kansas City       9         8          9           5        21,785
Dallas            8         6          6           6        23,573
San Francisco     3         1          1           8        26,696

(a) For asset data, see Annual Report of the Board of Governors (1990).
See also Meade and Sheets (2005a). GDP per capita (1990) was computed
from real GDP and population.


Rankings by region population and real output provide a somewhat different perspective on size. The former is computed by summing the population of all the counties in a Federal Reserve district and is thus precise; the latter is estimated by applying the share of a state's population to state-level output data for the 14 U.S. states that are assigned to more than one Federal Reserve district. Regions are ranked by population and output in 1990, which roughly corresponds to the middle of the voting sample. According to the population and output measures, San Francisco is the largest region, followed either by Atlanta or by New York depending on whether size is evaluated by population or output, respectively. It is worth noting that the San Francisco region accounts for about 20% of total population and output, while New York and Atlanta are about half as large.

Votes are generally not well aligned with region size regardless of the size measure used. The most egregious misalignment is for the San Francisco district, which ranks second for assets, first for population and output, but eighth for voting frequency. Atlanta is also clearly underweight with a voting rank of 10th. At the other extreme, the Boston and Kansas City districts have cast more votes than would be justified by any measure of their size.

Finally, Table 2 gives an estimate of Gross Domestic Product (GDP) per capita in 1990 for each Fed region based upon the population and output figures. The range of per capita income is relatively narrow--from a high of almost $30,000 in the New York region to a low of about $20,000 in the St. Louis district. This contrasts sharply with the very wide range of per capita income across members in Europe's monetary union, which is even wider when the prospective member countries in Central and Eastern Europe are taken into account.

FOMC participants dissent in the formal vote only very infrequently (Table 3). (2) Of the 4,067 votes cast by participants other than the chairman in FOMC meetings and conference calls between 1968 and 2004, only 6.9% (282 votes) were dissents. Board members dissent less frequently than bank presidents (6% and 8%, respectively), an empirical regularity noted in other studies that survives regardless of the time period examined. However, the magnitude of the difference in dissent rates depends on the sample period as the early 1980s was a time of relatively high dissent for board members. (3)
TABLE 3

FOMC Dissents, 1968-2004

                 Number  Percent

All voters         282     6.9
Board members      131     6.0
Bank presidents    151     8.0

Source: Meade and Sheets (2005a).


Direct examination of the FOMC's deliberations offers another perspective on the group's decision making. Blinder (2004) has characterized the Fed's policymaking body as an "autocratically collegial" committee, implying that the autocrat who heads the committee may act to suppress dissent from public view. Furthermore, Blinder and several coauthors have noted that "... Fed traditions dictate that a member should 'dissent' only if they find the majority's (that is, the chairman's) opinion unacceptable." (4) This observation suggests that the official dissent rates in Table 3 overstate the degree of consensus within the FOMC. It is possible to gain some insight into this apparent consensus by comparing dissenting votes with disagreements voiced by meeting participants during the discussion of Greenspan's policy proposal. Meade (2005) describes an original data set on voiced preferences (agreements and disagreements) of FOMC members using the meeting transcripts for the 1989-1997 period. (5) As Table 4 shows, the rate of disagreement with Greenspan's interest rate proposal during the 72 meetings between 1989 and 1997 was nearly 15%, (6) compared with a 7.5% dissent rate in the official vote. Furthermore, while both board members and bank presidents dissent more in the discussion than they do in the official vote, it is the bank presidents who are much more likely to voice disagreement during the deliberations. Voiced disagreements were the highest for the Kansas City, Richmond, and St. Louis. Not surprisingly, the New York district voiced practically no disagreements with Greenspan's policy proposal, which is consistent with the fact that Greenspan himself was appointed from the New York district, and the bank president from New York serves as the vice chairman of the FOMC and typically shows strong support for the chairman.
TABLE 4

Dissents Versus Disagreements, 1989-1997

                    Official Dissents      Voiced Disagreements

                   Voters  Board  Bank  Participants  Board  Bank

Total               55     24     31        178       32     146

Percent of votes     7.5    6.4    8.6       14.8      8.6    17.5
  Boston             6      6      6         16        1      15
  New York           0      0      0          3        0       3
  Philadelphia       1      0      1          5        1       4
  Cleveland (a)      8             8         15               15
  Richmond          10      4      6         28        7      21
  Atlanta (a)        0             0          9                9
  Chicago            6      6      0         18       10       8
  St. Louis          6      0      6         27        3      24
  Minneapolis (a)    3             3          8                8
  Kansas City       10      7      3         24        8      16
  Dallas             4      1      3          9        1       8
  San Francisco      1      0      1         16        1      15

(a) There was no board appointment from the district over this period.
Source: Meade and Sheets (2005a) and Meade (2005).


III. DATA, ESTIMATION, AND EMPIRICAL RESULTS

The disagreements voiced by FOMC policymakers declined after the FOMC decided to begin publishing lightly edited transcripts of its meetings. The data in Table 5 demonstrate this and show that the decline in voiced disagreement was particularly pronounced for policymakers who cast an official vote. The theoretical model of Meade and Stasavage (2008) predicts that once policymakers knew that their deliberations would eventually become public, they had an incentive to suppress their private information and go along with the consensus.
TABLE 5

Voiced Disagreement by Region

                     Voices Disagreement (Number)

                     1989-1992  1994-1997

By all participants      84         77
By official voters       48         27
  Boston                  6         10
  New York                2          1
  Philadelphia            4          1
  Cleveland              11          2
  Richmond                8         15
  Atlanta                 6          3
  Chicago                15          2
  St. Louis               7         17
  Minneapolis             2          6
  Kansas City             9         10
  Dallas                  8          0
  San Francisco           6         10

Source: Meade (2005).


When the disagreements voiced by FOMC policymakers are examined by Federal Reserve district, they exhibit no obvious systematic regional pattern. In six of the districts, voiced disagreements rose after 1993, while in the other six districts, voiced disagreements fell. Because Federal Reserve districts differ in size and voting power, the incentive to voice disagreement may vary.

In my empirical analysis, I examine the policy preferences voiced in FOMC meetings between 1989 and 1997 to see whether the effects of transcript publication identified by Meade and Stasavage (2008) are robust to the inclusion of regional variables. (7) Whether or not monetary policymakers are influenced by regional concerns has been the subject of two recent studies. Meade and Sheets (2005b) identified pronounced significant regional effects on FOMC voting behavior using the difference between the Federal Reserve district and the national unemployment rates as an explanatory variable in voting equations. Chappell, McGregor, and Vermilyea (2008) found evidence that the regional unemployment rate affects the policy preferences voiced by monetary policymakers during FOMC meetings. In the analysis here, part of the aim was to gather statistical evidence about a variety of regional proxies including, but not restricted to, a regional unemployment measure.

Throughout the empirical analysis, I exclude observations for all the 1993 meetings. FOMC participants found out at different times during that year that the literal transcriptions existed and thus may have determined at different points during 1993 that their comments would eventually become public. From reading the transcripts of 1993 meetings and conference calls, it is possible to conclude that Greenspan knew about the existence of the transcriptions as early as the end of 1992, other officials found out somewhat later, and some did not know until the Congressional hearings in October 1993. Therefore, the data sample includes a pretranscript period of 1989-1992 and a posttranscript period of 1994-1997, a total of 64 FOMC meetings.

The estimation equation has the following general form:

[VDISS.sub.it] = [[alpha].sub.0] + [[alpha].sub.1][TRANSCRIPT.sub.t] + [[alpha].sub.2][districtproxy.sub.it] + [C.sub.i][beta] + [X.sub.t][delta] + [[member of].sub.it],

where the dependent variable VDISS is a binomial indicator set equal to 0 in the case of voiced agreement and 1 in the case of disagreement with Greenspan's policy proposal for every FOMC participant (indexed by i) at each meeting (indexed by t) in the estimation sample period. The explanatory variables include a transcription dummy equal to unity after 1993 (TRANSCRIPT), a regional variable that represents different Federal Reserve districts (districtproxy), variables that capture policymaker characteristics (C), and variables that control for macroeconomic conditions (X).

District indicators tested include the average vote of each Federal Reserve district per FOMC meeting over the 1968-2004 period (VOTE), district real GDP in 1990 (GDP), the level of district per capita output in 1990 (PCGDP), district population in 1990 (POP), district bank assets in 1990 (ASSETS), and the absolute value of the difference between the district unemployment rate and the national unemployment rate in the month prior to each FOMC meeting (UNDIFF). (8) In addition to these district indicators, I also tested four variables based upon the ranking of each district in terms of real GDP, population, bank assets, and votes and three other variables based upon district percentages in total real GDP, population, and bank assets. Thus, a total of 12 district indicators were tested alternately in estimation.

There are five characteristic variables in each equation estimated: a dummy for voting bank presidents (BPVOTER), a dummy for nonvoting bank presidents (NonVOTER), and interaction terms between TRANSCRIPT and these dummy variables (Trans*BPVOTER and Trans*NonVOTER). (9) The final characteristic variable (EXPdiff), measured as the number of months on the FOMC relative to Greenspan, is intended to proxy for the rise in Greenspan's reputation in order to control for any effect this may have had on the willingness of other policymakers to disagree with him. In addition, each estimated equation includes three macroeconomic variables: the absolute value of real-time forecasts for consumer price inflation and the output gap (CPI and GAP, respectively) based upon the Fed staff's Greenbook projections and the Greenbook forecast error for productivity growth (PROD) that measures uncertainty about the economic situation.

Table 6 presents logit estimation results for a baseline Column (1) that excludes a district proxy and for another Column (2) that includes district-level fixed effects. Columns (3) through (8) of the table provide the estimation results for alternate specifications, estimated using standard logit techniques but including VOTE, GDP, PCGDP, POP, ASSETS, or UNDIFF, respectively, as the district indicator. Of the six district proxies shown in the table, three of them--the voting share in Column (3), per capita GDP in Column (5), and district bank assets in Column (7)--are negative and statistically significant. The negative parameter estimates for these variables point to lower voiced disagreement from officials in larger Federal Reserve districts, whether the size is measured by voting share, output per capita, or bank assets. This is perhaps not totally surprising in light of the fact that the New York district is both economically large and exercises an important weight in the overall vote. A likelihood ratio test strongly supports the inclusion of regional fixed effects in Equation (2), providing further support for a regional aspect. The other alternative district variables (GDP, POP, and UNDIFF) shown in the table, as well as most of the other district proxies tested, were not statistically significant. Of the district indicators tested but not shown in Table 6, only the district share of bank assets was significant and also entered with a negative sign. When the equation was reestimated using data for the entire 1989-1997 sample period (and the transcript dummy was excluded), the results were very similar; in that case too, VOTE, PCGDP, and ASSETS had a negative statistically significant effect on voiced disagreements, while the other regional proxies remained insignificant. Finally, it is interesting that the regional unemployment measure, UNDIFF, was insignificant. This is in stark contrast to the Meade and Sheets (2005b) finding that UNDIFF matters for FOMC voting behavior and the Chappell, McGregor, and Vermilyea (2008) finding that a different measure of regional unemployment matters for voiced preferences.
TABLE 6

Estimation Results (a)

Dependent Variable: VDISS

                         Voiced Agreement with Greenspan Interest
                        Rate Proposal (0), Voiced Disagreement (1)

                        Standard Logit Estimation Except as Noted


                           Baseline (1)   Fixed Effects Logit (2)

TRANSCRIPT               -1.15 ** (0.55)        -1.10 * (0.60)

District variables
  VOTE
  GDP
  PCGDP
  POP
  ASSETS
  UNDIFF

Characteristics
  BPVOTER                0.87 *** (0.34)       1.16 *** (0.36)
  Trans * BPVOTER            0.77 (0.61)           0.68 (0.62)
  NonVOTER                0.72 ** (0.33)       0.90 *** (0.34)
  Trans * NonVOTER       1.61 *** (0.58)       1.52 *** (0.59)
  EXPdiff                    2.23 (1.73)           2.40 (1.97)

Macro variables
  CPI                        0.06 (0.08)           0.06 (0.09)
  GAP                        0.02 (0.09)           0.01 (0.09)
  PROD                  -0.85 *** (0.22)      -0.87 *** (0.24)

Sample period           1989-1997 (excluding 1993)

Number of observations             1.068

                               Dependent Variable: VDISS

                         Voiced Agreement with Greenspan Interest
                        Rate Proposal (0), Voiced Disagreement (1)

                        Standard Logit Estimation Except as Noted

                                (3)               (4)

TRANSCRIPT               -1.16 ** (0.55)   -1.17 ** (0.56)

District variables
  VOTE                    -0.43 * (0.23)
  GDP                                          0.17 (0.33)
  PCGDP
  POP
  ASSETS
  UNDIFF

Characteristics
  BPVOTER                0.89 *** (0.34)    0.86 ** (0.34)
  Trans * BPVOTER            0.77 (0.61)       0.79 (0.61)
  NonVOTER                0.65 ** (0.33)    0.71 ** (0.33)
  Trans * NonVOTER       1.61 *** (0.58)   1.64 *** (0.59)
  EXPdiff                    2.21 (1.74)       2.21 (1.74)

Macro variables
  CPI                        0.05 (0.08)       0.06 (0.08)
  GAP                        0.01 (0.09)       0.02 (0.09)
  PROD                  -0.85 *** (0.22)  -0.85 *** (0.22)

Sample period

Number of observations

                               Dependent Variable: VDISS

                         Voiced Agreement with Greenspan Interest
                        Rate Proposal (0), Voiced Disagreement (1)

                        Standard Logit Estimation Except as Noted

                                (5)               (6)

TRANSCRIPT               -1.14 ** (0.55)   -1.19 ** (0.56)

District variables
  VOTE
  GDP
  PCGDP                  -0.08 ** (0.03)
  POP                                          7.44 (8.59)
  ASSETS
  UNDIFF

Characteristics
  BPVOTER                0.89 *** (0.36)   0.86 *** (0.36)
  Trans * BPVOTER            0.73 (0.61)       0.79 (0.61)
  NonVOTER                0.73 ** (0.33)    0.70 ** (0.33)
  Trans * NonVOTER       1.55 *** (0.58)   1.66 *** (0.59)
  EXPdiff                    2.65 (1.75)       2.21 (1.74)

Macro variables
  CPI                        0.05 (0.08)       0.06 (0.08)
  GAP                        0.01 (0.09)       0.02 (0.09)
  PROD                  -0.85 *** (0.22)  -0.85 *** (0.22)

Sample period

Number of observations

                               Dependent Variable: VDISS

                         Voiced Agreement with Greenspan Interest
                        Rate Proposal (0), Voiced Disagreement (1)

                        Standard Logit Estimation Except as Noted

                                (7)               (8)

TRANSCRIPT               -1.17 ** (0.55)   -1.16 ** (0.55)

District variables
  VOTE
  GDP
  PCGDP
  POP
  ASSETS                 -9.13 ** (3.54)
  UNDIFF                                      -0.03 (0.17)

Characteristics
  BPVOTER                1.04 *** (0.35)    0.86 ** (0.34)
  Trans * BPVOTER            0.75 (0.61)       0.78 (0.61)
  NonVOTER                0.73 ** (0.33)    0.72 ** (0.33)
  Trans * NonVOTER       1.56 *** (0.58)   1.61 *** (0.58)
  EXPdiff                    2.83 (1.73)       2.22 (1.73)

Macro variables
  CPI                        0.05 (0.08)       0.05 (0.08)
  GAP                        0.01 (0.09)       0.02 (0.09)
  PROD                  -0.85 *** (0.22)  -0.85 *** (0.22)

Sample period

Number of observations

Notes: Robust standard errors are in parentheses. EXPDIFF, GDP, PCGDP,
POP, and ASSETS divided by 1,000 to aid presentation of results.
(a) Constants included but not reported.
* Statistical significance at the 10% level; ** statistical
significance at the 5% level; *** statistical significance at the 1%
level.


Across Columns (3)-(8), the parameter estimates for the transcript dummy, the characteristic variables, and the macro variables are generally quite similar to those for Column (1). The transcription dummy is consistently negative and statistically significant, pointing to an effect of known transcription even after other factors, including economic conditions, are taken into account. The variable intended to proxy for macroeconomic uncertainty, PROD, is also significant and indicates greater voiced agreement with Greenspan's proposal as uncertainty rises. The real-time forecasts for inflation and the output gap are not significant, and neither is the proxy for Greenspan's reputation.

With respect to the voter characteristics, the dummy variables for voting and nonvoting bank presidents (BPVOTER and NonVOTER, respectively) indicate that those policymakers are significantly more likely to voice disagreement with a given policy proposal than board members. While the interaction term Trans*BPVOTER was not significantly different from 0 in any of the equations estimated, Trans*NonVOTER was positive and highly significant in all the specifications. Taking these results together with the magnitude of the transcription dummy shows that voting bank presidents are more likely to voice disagreement than board members but that disagreements drop for both groups after 1993. In contrast, bank presidents who do not cast an official vote show a rise in disagreement after 1993 because the NonVOTER dummy and the interaction term more than offset the estimated negative transcription effect.

Overall, the results show that the transcription effects are robust, statistically important, and invariant to the inclusion of regional indicators. Furthermore, the regional indicators that are significant uniformly indicate that larger districts are more likely to agree with the Chairman's proposal, while smaller districts are more likely to voice dissent. Although many of the regional proxies were statistically insignificant, three of the most important indicators--the voting share, regional per capita output, and district bank assets--are highly significant and suggest that there is some regional aspect to voiced policy preferences. Additional support for this is provided from fixed effects estimation. Thus, this article provides some evidence for regional effects on the pattern of disagreements voiced during FOMC deliberations and establishes that the transcription effect remains robust to their inclusion.

IV. CONCLUSIONS

This article expands the work on transcription and FOMC deliberations to consider the possibility of a regional dimension to disagreements voiced at FOMC meetings before and after 1993. The major results in this article are that the effects of transcript publication on FOMC deliberation noted by Meade and Stasavage (2008) are robust to the inclusion of regional variables and that there is some evidence that regional effects (as proxied by the district voting share, district per capital output, district bank assets, and logit estimation with district fixed effects) matter--that is, that larger Federal Reserve districts are less likely to voice disagreement with a given interest rate proposal than smaller districts. The article adds to a small but growing literature on the regional aspect of Federal Reserve deliberation and debate. A remaining question perhaps worthy of additional study is why disagreements voiced by nonvoting policymakers rise, rather than fall, after the decision was made in 1993 to publish the meeting transcripts. The different behavior of disagreements for voting and nonvoting policymakers suggests that the two groups face different incentives during meeting discussions.

REFERENCES

Annual Report of the Board of Governors of the Federal Reserve System. Washington, DC: Board of Governors of the Federal Reserve System, 1990.

Belden, S. "Policy Preferences of FOMC Members as Revealed by Dissenting Votes." Journal of Money, Credit, and Banking, 21, 1989, 432-41.

Blinder, A. S. Central Banking in Theory and Practice. Cambridge, MA: MIT Press, 1998.

--The Quiet Revolution. New Haven, CT: Yale University Press, 2004.

Blinder, A., C. Goodhart, P. Hildebrand, D. Lipton, and C. Wyplosz. How Do Central Banks Talk? Geneva Reports on the World Economy Number 3. London: Centre for Economic Policy Research, 2001.

Chapped, H. W. Jr., R. R. McGregor, and T. A. Vermilyea. "Regional Economic Conditions and Monetary Policy." European Journal of Political Economy, 24, 2008, 283-93.

Geraats, P. "Central Bank Transparency." Economic Journal, 112, 2002, 532-65.

Gersbach, H., and V. Hahn. "Voting Transparency, Conflicting Interests, and the Appointment of Central Bankers." Economics & Politics, 16, 2004, 321-45.

--"Voting Transparency in a Monetary Union,." CEPR Discussion Paper 5155, 2005.

--"Should the Individual Voting Records of Central Bankers Be Published?" Social Choice and Welfare, 30, 2008, 655-83.

Gildea, J. A. "Explaining FOMC Members' Votes," in The Political Economy of American Monetary Policy, edited by T. Mayer. Cambridge, MA: Cambridge University Press, 1990, 211-28.

--"The Regional Representation of Federal Reserve Bank Presidents." Journal of Money, Credit, and Banking, 24, 1992, 215-25.

Havrilesky, T. M., and R. L. Schweitzer. "A Theory of FOMC Dissent Voting with Evidence from the Time Series," in The Political Economy of American Monetary Policy, edited by T. Mayer. Cambridge, MA: Cambridge University Press, 1990, 197-210.

Meade, E. E. "The FOMC: Preferences, Voting, and Consensus." Federal Reserve Bank of St. Louis Review, 87(2), 2005, 93-101.

Meade, E. E., and D. N. Sheets. "Documenting FOMC Voting Patterns," Workshop Proceedings No. 7, Vienna: Austrian National Bank, 2005a.

--"Regional Influences on FOMC Voting Patterns." Journal of Money, Credit, and Banking, 37, 2005b, 661-77.

Meade, E. E., and D. Stasavage. "Publicity of Debate and the Incentive to Dissent: Evidence from the US Federal Reserve." Economic Journal, 118, 2008, 695-717.

Meyer, L. H. A Term at the Fed: An Insider's View. New York: Harper Collins, 2004.

Ottaviani, M., and P. Sorensen. "Information Aggregation in Debate: Who Should Speak First?" Journal of Public Economics, 81, 2001, 393-421.

Purposes and Functions. Washington, DC: Board of Governors of the Federal Reserve System, 2005.

(1.) A 1942 amendment to the Federal Reserve Act specifies that one vote is shared by the following districts: Chicago and Cleveland; Boston, Philadelphia, and Richmond; Atlanta. Dallas, and St. Louis; Kansas City, Minneapolis, and San Francisco.

(2.) For a discussion of possible explanations for the low dissent rate, including the structure of the FOMC and the power of its chairman, see Blinder (1998, 2004).

(3.) See Meade and Sheets (2005a) for voting statistics computed over different periods and for information on dissents by Federal Reserve district.

(4.) See Blinder. Goodhart, Hildebrand, Lipton, and Wyplosz (2001, p. 39).

(5.) Greenspan spoke first and put a policy proposal on the table during FOMC discussions of monetary policy. He then called on other policymakers (in no fixed order) to speak. Thus, voiced preferences record the views expressed about Greenspan's proposal. The official vote was recorded at the end of the discussions, with Greenspan casting his vote first. Meyer (2004) provides additional detail on the structure of FOMC meetings.

(6.) Figures in Table 4 include views voiced by voting and nonvoting bank presidents. Nonvoting bank presidents disagree more often than voting ones do (19.3% and 15.2%, respectively).

(7.) The 1989-1997 sample period corresponds to that of Meade and Stasavage (2008).

(8.) VOTE is reported on Table 1. PCGDP is reported on Table 2. SIZE, POP, and ASSETS are the data that underlie the rankings on Table 2. UNDIFF is the variable used in Meade and Sheets (2005b).

(9.) The omitted category is board members. There is some variation across time in the bank president voter and nonvoter categories since the vote rotates annually.

ABBREVIATIONS

FOMC: Federal Open Market Committee

GDP: Gross Domestic Product

ELLEN E. MEADE *

*A much earlier version of this article entitled "Dissents and Disagreements in the Fed's FOMC: Understanding Regional Affiliations and Limits to Transparency" was prepared for the workshop "Central Banking by Committee" organized by The Netherlands Bank and presented at the 2006 Western Economic Association International Conference. I would like to thank an anonymous referee and participants at The Netherlands Bank conference and Western Economic Association International session for useful comments and Robert Kahn and Nathan Sheets for helpful discussions.

Meade: Associate Professor, Department of Economics, American University, 4400 Massachusetts Avenue NW, Washington, DC 20016. Phone 1-202-885-3997. Fax 1-202-885-3790, E-mail meade@american.edu

doi: 10.1111/j.1465-7287.2009.00175.x
COPYRIGHT 2010 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Meade, Ellen E.
Publication:Contemporary Economic Policy
Article Type:Report
Geographic Code:1USA
Date:Apr 1, 2010
Words:5405
Previous Article:Casinos, hotels, and crime.
Next Article:Is tighter fiscal policy expansionary under fiscal dominance?: hypercrowding out in Latin America.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters