THREE ESSAYS ON MONETARY POLICY

By

Matthew Schaffer

A DISSERTATION

Michigan State University

in partial fulfillment of the requirements

Submitted to

for the degree of

Economics – Doctor of Philosophy

2018

ABSTRACT

THREE ESSAYS ON MONETARY POLICY

By

Matthew Schaffer

This dissertation contains three empirical studies on monetary policy.

The first chapter investigates the impact of banking deregulation on the effectiveness of monetary
policy and provides new evidence on how bank-level heterogeneity affects the lending channel of
transmission. Exploiting the staggered deregulation of interstate banking in the U.S. throughout
the 1980’s, I find that the deregulation strengthens the effect of monetary policy on bank lending,
doubling the response of loan growth to monetary shocks. This effect occurs primarily for small
and relatively illiquid banks, pointing to a strengthening of the bank lending channel. After
deregulation this subset of banks engages in a larger substitution of securities for bank loans
following a contractionary monetary shock. Changes in bank market structure and loan portfolio
composition cannot explain these effects of the deregulation. By contrast, the findings point to a
dilution in the strength of bank-borrower customer relationships and a stronger propensity of banks
to cut loans to their customers.

The second chapter studies the response of stock prices to monetary policy, distinguishing effects
of exogenous shocks from “Delphic" shocks that reveal the Federal Reserve’s macroeconomic
forecasts. A measure of Federal Reserve private information that exploits differences in central
bank and market forecasts is used to decompose monetary policy surprises into these separate
components. Contractionary policy shocks of either type lower stock prices with exogenous shocks
having a larger negative effect. There is also some evidence of an asymmetry; when FOMC
meetings are unscheduled or when the fed funds rate reverses direction, stock prices actually rise
in response to a contractionary Delphic shock.

The third chapter examines the response of real personal income in eight United States regions
to monetary policy shocks using a structural VAR framework. The external instruments approach

is used for identification. I split the sample into two periods 1958:Q1 – 1992:Q4 and 1993:Q1-
2015:Q2 and find markedly different responses between the two. In the early period real personal
income decreases in all regions following a contractionary shock while in the later period little
response is seen. I investigate two potential reasons why there has been a reduction in differential
regional responses to monetary policy over time. The evidence suggests that the reduction may be
associated with a homogenization in regional industry composition over recent decades.

To my parents, Dan and Rhonda Schaffer, for their constant love and support.

iv

ACKNOWLEDGEMENTS

I would like to thank the members of my committee, especially Professor Raoul Minetti for his
invaluable advice, guidance, and support. I would also like to thank Professor Aeimit Lakdawala
for serving as an irreplaceable mentor the last three years.

Additionally I would like to thank my colleagues at Michigan State, particularly Obafemi

Elegbede, without whom I would not have made it very far.

v

TABLE OF CONTENTS

LIST OF TABLES .
LIST OF FIGURES .
CHAPTER 1 BANK REGULATION AND MONETARY POLICY TRANSMISSION:

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
x
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.

.

.

.

.

.

.

.

.
.

.
.
.

. .

. .

Introduction .

1.4 Results .

1
EVIDENCE FROM THE U.S. STATES LIBERALIZATION . . . . . . . .
1
1.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
1.2 Geographic Banking Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Data and Estimation .
5
1.3.1 Monetary Policy Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
1.3.2 Banking Data .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
1.3.3 Estimation .
9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
. .
.
1.4.1
Impact of Deregulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4.2 Role of Bank Lending Channel . . . . . . . . . . . . . . . . . . . . . . . . 15
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.1 Bank Market Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.2 Loan Portfolio Composition . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.5.3 Bank-Borrower Relationships . . . . . . . . . . . . . . . . . . . . . . . . . 23
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

.
1.6 Aggregate Effects .
.
1.7 Conclusion .
.
. .
APPENDIX .

1.5 Potential Explanations .

.
.
.
.
. .

. .
. .

. .

.

CHAPTER 2 FEDERAL RESERVE PRIVATE INFORMATION AND THE STOCK

.

.

.

.
.

.
.

.
.

.
.

. .

. .

. .

2.3 Data .

Introduction .

MARKET .
. .
.

2.1
2.2 Stock Prices and Monetary Policy . . . . . . . . . . . . . . . . . . . . . . . . .

Stock Price Response to Exogenous and Delphic Shocks
.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
. 58
2.2.1 Monetary Policy Surprise from Futures Data . . . . . . . . . . . . . . . . . 58
2.2.2
. . . . . . . . . . 62
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.3.1 Monetary Policy Surprise . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.3.2
Federal Reserve Private Information . . . . . . . . . . . . . . . . . . . . . 65
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.4.1
. . . . . . . . . . . . . . . . . 67
2.4.2 Monetary Policy Surprise and Private Information . . . . . . . . . . . . . . 69
2.4.3
. . . . . . . . . . 71
2.4.4 VAR Based Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . 75
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Stock Price Response to Exogenous and Delphic Shocks

Stock Prices and Monetary Policy Surprise

2.5 Conclusion .
APPENDIX .

2.4 Results .

. .
. .

.
.
. .

.
.
. .

. .

. .

. .

.
.

.

.

.

CHAPTER 3 REGIONAL RESPONSES TO MONETARY POLICY OVER TIME . . . . 97

vi

. .
. .

.
.

.
.
.
.
.
.
.
.
. .
.
.

Introduction . .

. .

.
3.1
3.2 Previous Literature .
.
3.3 Estimation . .
3.4 Results . .
. .
.
Full Sample .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 102
. 103
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 103
3.5 Regional Monetary Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.6 Regional Industry Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.7 Conclusion . .
APPENDIX . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

3.4.1
3.4.2 Early Sample .
3.4.3 Recent Sample

. .
. .

.
.
. .

.
.

BIBLIOGRAPHY .

.

.

.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

vii

LIST OF TABLES

Table 1.1: Timing of State Deregulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Table 1.2: Summary Statistics - All Banks

. . . . . . . . . . . . . . . . . . . . . . . . . . 30

Table 1.3: Summary Statistics - by Bank Size . . . . . . . . . . . . . . . . . . . . . . . . . 31

Table 1.4:

Impact of Deregulation on Lending Response to Monetary Policy . . . . . . . . 32

Table 1.5:

Impact on Lending Response: All Coefficients - Baseline Model . . . . . . . . . 33

Table 1.6: Robustness: NBR Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Table 1.7:

Impact on Loan Pricing Response to Monetary Policy . . . . . . . . . . . . . . . 35

Table 1.8:

Impact on Lending Response: by Bank Size . . . . . . . . . . . . . . . . . . . . 36

Table 1.9:

Impact on Loan Pricing: by Bank Size . . . . . . . . . . . . . . . . . . . . . . . 37

Table 1.10: Impact on Lending Response: by Bank Liquidity . . . . . . . . . . . . . . . . . 38

Table 1.11: Impact on Lending Response: by Bank Capitalization . . . . . . . . . . . . . . . 39

Table 1.12: Market Structure .

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Table 1.13: Impact on Lending Response: by Bank Market Power . . . . . . . . . . . . . . . 41

Table 1.14: Impact on Lending Response: by Market Concentration . . . . . . . . . . . . . 42

Table 1.15: Loan Portfolio Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Table 1.16: Impact on Lending Response: by BHC Affiliation . . . . . . . . . . . . . . . . . 44

Table 1.17: Impact on Lending Response: by Size, Liquidity, and BHC Affiliation . . . . . . 45

Table 1.18: Balance Sheet Adjustment: by Size, Liquidity, and BHC Affiliation . . . . . . . 46

Table 1.19: Share of National Lending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Table 1.20: Impact on Loan Growth Response: State-Level Results . . . . . . . . . . . . . . 48

Table 2.1: FOMC, Greenbook & Blue Chip Dates

. . . . . . . . . . . . . . . . . . . . . . 82

viii

Table 2.2: Summary Statistics - Monetary Policy Surprise and S&P 500 . . . . . . . . . .

. 83

Table 2.3: Stock Market Response to Monetary Policy Surprise - 30 Minute Window . . . . 84

Table 2.4: Stock Market Response to Monetary Policy Surprise - Daily Window . . . . . . 85

Table 2.5: 1st Step - Monetary Policy Surprise on Private Information . . . . . . . . . . .

. 86

Table 2.6: Summary Statistics - Exogenous and Delphic Shocks . . . . . . . . . . . . . . . 87

Table 2.7: 2nd Step - Response of Stock Prices to Exogenous and Delphic Shocks

. . . .

. 88

Table 2.8: 2nd Step - Subsamples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 89

Table 2.9: Excess Stock Return Variance Decomposition . . . . . . . . . . . . . . . . . . . 90

Table 2.10: Response of Excess Returns to Exogenous & Delphic Shocks . . . . . . . . . . . 91

Table 2.11: Response of Excess Returns on Unscheduled/Turning Point Meetings

. . . . . . 92

Table 3.1: State-Level Monetary Policy Channels . . . . . . . . . . . . . . . . . . . . . . . 112

Table 3.2: State-Level Monetary Policy Response and Industry Shares . . . . . . . . . . . . 113

Table 3.3: State Industry Share Standard Deviations

. . . . . . . . . . . . . . . . . . . .

. 114

ix

LIST OF FIGURES

Figure 1.1: Monetary Policy Shock Series . . . . . . . . . . . . . . . . . . . . . . . . . .

. 49

Figure 1.2: Real Loan Growth .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Figure 1.3: Bank Credit to Total Private U.S. Credit

. . . . . . . . . . . . . . . . . . . . . 51

Figure 1.4: Average Loan Rate .

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Figure 1.5: Bank Type Share of National Lending . . . . . . . . . . . . . . . . . . . . . . 53

Figure 2.1: Monetary Policy Surprise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Figure 2.2: Private Information Variables . . . . . . . . . . . . . . . . . . . . . . . . . .

. 94

Figure 2.3: Stock Returns vs Monetary Policy Surprises . . . . . . . . . . . . . . . . . .

. 95

Figure 2.4: Exogenous and Delphic Shocks . . . . . . . . . . . . . . . . . . . . . . . . .

. 96

Figure 3.1: Cumulative Response to Monetary Policy Shock (1958.Q1-2015.Q2) . . . . . . 115

Figure 3.2: Cumulative Response to Monetary Policy Shock (1958.Q1-1992.Q4) . . . . . . 116

Figure 3.3: Cumulative Response to Monetary Policy Shock (1993.Q1-2015.Q2) . . . . . . 117

Figure 3.4: Cumulative Response to Monetary Policy Shock (1984.Q1-2008.Q4) . . . . . . 118

x

CHAPTER 1

BANK REGULATION AND MONETARY POLICY TRANSMISSION: EVIDENCE

FROM THE U.S. STATES LIBERALIZATION

1.1 Introduction

The impact of regulatory changes in recent decades on the U.S. banking industry has been
well documented.1 A relatively unexplored aspect of these regulatory changes has been their
effect on monetary policy. As a recent literature documents, structural features of the banking
sector can influence the responsiveness of lending to monetary policy.2 Bank lending is an
important mechanism through which monetary policy is transmitted to the broader economy. For
both policymakers and bank regulators it is therefore important to know how changes in banking
regulation may impact the effectiveness of monetary policy. In this paper we address these issues
by exploiting the natural experiment provided by the staggered state-level removal of geographic
banking restrictions throughout the 1980’s.

We find that lending becomes more responsive to monetary policy after a bank’s home state
removes interstate geographic banking restrictions.3 After a state removes interstate restrictions
the response of real lending growth to monetary policy doubles, however the removal of intrastate
restrictions has no effect. More precisely, our results indicate that interstate banking deregulation
strengthens the bank lending channel of monetary policy transmission. Kashyap and Stein (2000)
establish that the bank lending channel is operative for small and relatively illiquid banks. We find
that interstate deregulation increases the sensitivity of lending to monetary policy primarily for
1See Jayaratne and Strahan (1998), Berger, Demsetz, and Strahan (1999), Berger and DeYoung
(2001), Stiroh and Strahan (2003), and Berger, Demirguc-Kunt, Levine, and Haubrich (2004)
among others.

2See Adams and Amel (2011), Olivero, Li, and Jeon (2011), and Amidu and Wolfe (2013).
3There are two main types of restrictions: those on interstate (out-of-state) banking and those
Interstate deregulation allows out-of-state bank holding
on intrastate (within-state) branching.
companies to acquire and operate in-state banks. Intrastate deregulation allows banks headquartered
within a state to open additional branches through mergers and acquisitions.

1

small banks and that the effect is larger for relatively illiquid banks, pointing to a strengthening of
the bank lending channel.

We consider a variety of possible explanations for the greater impact of monetary policy after
interstate deregulation. Deregulation increases average bank market power, increases local banking
concentration, and decreases state banking concentration. Amidu and Wolfe (2013), Yang and
Shao (2016), and Adams and Amel (2011) find that lending is more responsive to monetary policy
for banks with greater market power or banks located in more highly concentrated local markets.
However, these changes in bank market structure cannot account for the effect of deregulation,
as we find that banks with greater market power are less responsive to monetary policy, and that
concentration has no impact on the relationship between policy and lending. Den Haan, Sumner,
and Yamashiro (2007) find that commercial and industrial loans increase following a contractionary
monetary policy shock while real estate and consumer loans decrease. We find all three types of
loans become more sensitive to monetary policy after deregulation at roughly the same magnitude,
ruling out the possibility that deregulation’s impact on monetary policy is driven by changes in the
composition of bank loan portfolios.

One mechanism that arises as a potential candidate for rationalizing our findings is a change in
the intensity of bank-borrower relationships and the accompanying propensity of banks to shield
their customers from negative shocks. Ashcraft (2006) finds that stand alone banks are more
responsive to monetary policy than banks affiliated with a holding company. We find this is true
prior to interstate deregulation, but that affiliated banks actually become more responsive post-
deregulation. After the removal of interstate restrictions small banks affiliated with a holding
company are unique in responding to contractionary monetary policy by more strongly adjusting
the asset side of their balance sheets towards securities and away from lending. The literature has
found that bank-customer relationships can play an important role in the availability of loans and
in cushioning firms from contractionary shocks. However their strength can be diluted by bank
mergers and acquisitions, especially when these entail consolidation across geographically distant

2

markets.4 One explanation for our results is that banks affiliated with holding companies engage
in more transactional lending than relationship lending. After the deregulation and associated
consolidation process these banks become more inclined to curtail loans in response to an adverse
monetary shock.5

Finally, we investigate the impact of deregulation for loan growth at the aggregated state-level.
The effect of interstate deregulation on the response of total loan growth to monetary policy is
negative but insignificant. However the effect on aggregate lending from a subsample of small
banks affiliated with bank holding companies is negative, significant, and relatively large at -8%.
These banks make up 16% of total lending on average, hence there is a non-negligible effect on
total lending at the state-level.

The rest of the paper unfolds as follows. Section 1.2 details a brief history of geographic
banking regulation in the United States. Section 2.3 discusses the data and estimation. Section 1.4
presents the main results and investigates the role of the bank lending channel. Section 1.5 explores
potential explanations. Section 1.6 documents the effect of deregulation for aggregate lending at
the state-level, and section 2.5 concludes.

1.2 Geographic Banking Regulation

Since the 19th century most U.S. states have imposed restrictions on the ability of banks to
expand geographically 6. These restrictions typically included an outright ban on out-of-state banks
owning in-state banks as well as strict limitations on the number of branches that an in-state bank
can operate. Deregulation of these restrictions took place in the majority of states from the mid-
1970’s to mid-1990’s. Over this time frame every state other than Hawaii began to allow interstate
banking and 35 different states removed restrictions on intrastate branching 7.

4 See Berger and Udell (1995), Berger and Udell (2002), Degryse and Ongena (2005), Minetti

(2011), and Araujo and Minetti (2011).

5See Peek and Rosengren (1998), Calomiris and Karceski (2000), Sapienza (2002), Bofondi and
Gobbi (2006), and Berger and Bouwman (2009) on the effect of mergers, entry, and organizational
structure on lending practices.

6See Kroszner and Strahan (1999) for a detailed history
714 states already allowed intrastate branching and one, Iowa, did not deregulate at all

3

Interstate banking was effectively banned by the Douglas amendment to the Bank Holding
Company Act of 1956. The amendment stated that a bank holding company (BHC) could not
acquire an out-of-state bank unless the state the bank is located in has passed a statue explicitly
allowing such transactions. Maine was the first state to pass such a statue and began allowing
out-of-state bank holding companies to acquire Maine banks in 1978. Deregulation picked up in
the 1980’s, particularly after passage of the federal Garn-St Germain Act of 1982, which amended
the Bank Holding Company Act to allow out-of-state bank holding companies to acquire failed
banks or thrifts in any other state. States began entering reciprocal regional or national agreements
through which bank holding companies in any state which had agreed to the arrangement could
purchase banks operating in any of the other states. 38 states joined such an agreement between
1984 and 1988 8.

Restrictions on intrastate branching were often removed in three steps. First, BHC’s would
be allowed to own multiple banks within one state, with each subsidiary operating as a separate
institution - e.g., a depositor at one subsidiary could not access funds at a different subsidiary.
Second, banks were allowed to establish additional branches through mergers and acquisitions
(M&A). Importantly, this allowed BHC’s operating within a state to convert their subsidiaries into
branches of a single bank. Finally, unrestricted branching was allowed in which banks were free
to open new within-state branches as they pleased. The literature has focused on the second step,
allowing branching via M&A, as the most important one. Most states had removed restrictions on
in-state BHC expansion by the mid-1970’s. Of the 15 who removed such restrictions after 1975, this
often occurred around the same time that M&A branching restrictions were removed. Similarly,
most states allowed unrestricted branching only a short time after allowing M&A branching.

Table 1.1 lists the year for which each state and the District of Columbia began to allow
branching via M&A and interstate banking 9. Congress passed the Riegle-Neal Interstate Banking
and Branching Efficiency Act in 1994, which allowed for national interstate banking and branching,

8Amel (1993)
9Dates from Amel (1993) and Kroszner and Strahan (1999).

4

effectively ending the period of state-level deregulation. The legislation went fully into effect
in 1997 but many states adopted early in mid-1995. Thus the period of interest for state-level
deregulation is from 1976 (when U.S. bank-level data becomes available) to 1994.

1.3 Data and Estimation

1.3.1 Monetary Policy Data

Conventional measures of monetary policy, such as the federal funds rate, can be problematic
for two reasons. First, such measures are endogenous, i.e.
they change in response to past and
contemporaneous economic conditions. Second and more importantly, they reflect anticipatory
movements by the monetary policymaker. Romer and Romer (2004) seek to surmount these issues
by devising a new series of monetary policy shocks. First, they construct a series of intended
federal funds rate changes around FOMC meetings by combining information from the Weekly
Report of the Manger of Open Market Operations and narrative accounts of each FOMC meeting.
Second, using the Fed’s internal Greenbook forecasts, they purge the series of variation attributable
to forecasts of future macroeconomic activity through the following regression:

2

ϕ(cid:101)πmi +

i=−1

2

γi∆(cid:101)ymi +
λi(∆(cid:101)ymi − ∆(cid:101)ym−1,i)
2

θ((cid:101)πmi −(cid:101)πm−1,i) + ρ(cid:101)um0 + 
m

i=−1

i=−1

2


i=−1

+

∆ f fm = α + β f f bm +

(1.1)

where ∆ f fm is the change in the intended federal funds rate at meeting m, f f bm is the level of the

intended funds rate prior to meeting m, ∆(cid:101)y is forecasted real output growth,(cid:101)π is forecasted inflation,
and(cid:101)u is the forecasted unemployment rate. Note that the previous period and contemporaneous
10. The residual of the above estimated equation,(cid:98)
m, then becomes a cleaner measure of monetary

forecasts of output growth and inflation are included in addition to forecasts of the next two quarters

policy shocks purged of endogenous and anticipatory variation. This measure will be our baseline
indicator of monetary policy and will henceforth be referred to as the RR shock series.

10The previous period forecasts are typically observed data.

5

Using the change in the effective fed funds rate as a measure of monetary policy may be
particularly troublesome when estimating the effect of monetary policy on bank lending. Bluedorn,
Bowdler, and Kochc (2017) provide evidence that such estimates at the individual bank-level are
quite sensitive to the measure of monetary policy used. They argue that the RR shocks eliminate
some of the major sources of endogenous variation plaguing other measures of policy. For instance,
suppose the FOMC increases the fed funds rate due to anticipated higher output growth in the coming
quarters. Higher output growth is likely to be associated with an increased demand for bank loans.
A regression of the change in bank lending on lagged changes in the fed funds rate may therefore
show a positive correlation, i.e. that contractionary monetary policy is associated with increased
bank lending. Such an increase in the fed funds rate will not show up in the RR shocks however,
mitigating such endogeneity concerns.

We use an updated series of RR shocks from Coibion, Gorodnichenko, Kueng, and Silvia
(2012). The series is initially calculated at the frequency of FOMC meetings then aggregated to a
quarterly average. The updated RR shocks as well as the change in the effective fed funds rate are
plotted in Figure 1.1. The RR shock is smaller in magnitude than the change in the fed funds rate,
which is unsurprising given that it is a residual of the latter. The two series typically move together
and have a high positive correlation of 0.82. There is a noticeable period of outliers for both series
from 1979 to 1982. During this period the Federal Reserve was targeting non-borrowed reserves
(NBR) rather than the fed funds rate which resulted in large and volatile gyrations in the funds rate.
Our baseline specifications include year or quarter dummy variables to account for this period.

1.3.2 Banking Data

Bank-level data is from the Consolidated Reports of Condition and Income ("Call Reports") which
all banks in the U.S. are required to file on a quarterly basis with the Federal Financial Institutions
Examinations Council (FFIEC). We follow Kashyap and Stein (2000) and Den Haan, Sumner,
and Yamashiro (2002) in defining our sample as all commercial banks which are insured, have
positive assets, and are located in the fifty states or Washington, D.C. Since mergers typically

6

create discontinuities in the acquiring bank’s balance sheet, a bank observation is dropped from
the sample in any quarter in which a merger occurs. To prevent outliers from driving the results
a bank-quarter is dropped whenever total loan growth is more than five standard deviations away
from that quarter’s average loan growth. Additionally, a bank-quarter is dropped if there are not
four preceding quarterly observations for total loan growth. This leaves slightly over 900,000
observations from 16,000 different banks in the sample.

Summary statistics for bank-level variables of interest are reported in table 1.2. The first two
columns show summary statistics for the entire sample (1976Q2-1994Q4). The third and fourth
columns show summary statistics for the early part of the sample (1976Q2-1985Q4) when a majority
of states had not deregulated. The fifth and sixth columns show summary statistics for the later part
of the sample (1986Q1-1994Q4) when a majority of states had deregulated. The main bank-level
variable of interest is real loan growth11. Over the entire sample average quarterly loan growth at
a single bank is 1.13% with a standard deviation of 7.25%. Average real loan growth across all
banks is plotted in figure 1.2. The series is relatively stable across the sample except for the period
of NBR targeting in the early 1980’s, which features a sharp drop. The share of total U.S. credit
included in our sample is substantial. Figure 1.3 plots aggregated commercial bank lending in our
data as a share of total private credit in the U.S. Over this time period commercial bank lending in
our sample accounts for 30-43% of all private sector credit.

The Call Reports do not directly include data on loan rates. However, following Jayaratne and
Strahan (1998) and Zarutskie (2013), a proxy for the average interest rate on a bank’s loan portfolio
can be calculated as total interest and fee income on loans divided by total loans12. Interest and
fee income on loans is reported on a year to date basis. Hence, the previous quarter’s value is
subtracted from the current value to obtain a quarterly measure. Interest and fee income on loans
is reported biannually prior to 1983. In order to use our full sample we replace the missing first
quarter observations with half of the second quarter value and the missing third quarter observations

11Call report loan data is in nominal terms; we adjust for inflation using CPI.
12Jayaratne and Strahan (1998) uses this approach to study the effects of deregulation on loan

pricing. Zarutskie (2013) studies the effects of deregulation and securitization on loan pricing.

7

with the average of the second and fourth quarter values. All results below are robust to leaving
the missing values empty however. The annualized mean of a bank’s average loan rate is roughly
11.5% for the entire sample, with a standard deviation just over 4%. The average loan rate across
all banks is plotted in figure 1.4. As with real loan growth, there are large variations in the early
1980’s before stabilizing for the rest of the sample.

Loan growth for the three major loan categories are included as well. Real estate lending saw
the largest average growth over the sample at 2.16% per quarter. Commercial and industrial (C&I)
lending growth averaged 0.76% for the entire sample, and saw a large drop from 1.63% in the early
part of the sample to -0.33% in the later part. Similarly. consumer lending grew an average of
1.04% in the early part of the sample before falling to -0.27% in the later part. Each categories
average share of total lending reflects these growth trends as the share of real estate lending grew
over the sample while the share of C&I and consumer lending decreased.13

Other bank-level variables of interest include total assets, security holdings, liquidity ratio,
equity ratio, and bank holding company affiliation. Average bank assets almost double from the
early part of the sample to the later part, with a mean of $173 million for the entire sample.
We follow Kashyap and Stein (2000) in defining our securities variable. There is not a consistent
variable tracking securities in the Call Reports over the entire sample. Prior to 1984 total securities is
calculated as the sum of U.S. Treasury securities, U.S. government agency and corporate obligations,
obligations of states and political subdivisions, all other bonds, stocks, and securities, and fed funds
sold and securities purchased under agreements to resell. From 1984 to 1993 it is calculated as
the sum of the book value of total investment securities, assets held in trading accounts, and fed
funds sold. A consistent definition is not available for 1994, the final year of our sample. Average
security holdings in a quarter double from the early part of the sample to the later part, with an
overall average of $38 million. Liquidity ratio is defined as the ratio of cash and reserves to total
liabilities. The average liquidity ratio for the sample is 0.09. Equity ratio is measured as the
ratio of total equity to total assets and is stable across the sample with a mean of 0.09. Finally,

13Zarutskie (2013) studies these trends in detail.

8

affiliation with a bank holding company increases greatly over this time frame, as restrictions on
bank expansion and acquisition are removed. Three measures of bank market structure are reported:
the Lerner Index, which measures a bank’s market power, county-level HHI, which measures local
banking concentration, and state-level HHI, which measures state banking concentration. All three
measures increase from the early to late part of the sample.

Summary statistics split by bank size are reported in table 1.3. As is conventional in the
literature, small banks are defined as any bank under the 95th percentile in total assets for a given
quarter. Large banks are defined as any bank above the 95th percentile for a given quarter.14
Average assets for small banks over the entire sample is $51 milllion, whereas average assets for
large banks is much larger at $2.5 billion. Loan growth is higher on average for small banks at
1.15% versus 0.86% for large banks. Average loan rates are similar for both. Real estate lending
makes up the largest share of loans for both although it is a higher share for small banks. Large
banks have higher liquidity ratios and lower equity ratios on average, and are more likely to be
affiliated with a bank holding company. Large banks also have higher average market power.

1.3.3 Estimation

How did deregulation of geographic banking restrictions impact the effectiveness of monetary
policy? In answering this question we follow the literature in specifying two distinct types of
deregulation: intrastate branching deregulation and interstate banking deregulation. To determine
the effect of each type of deregulation on monetary transmission we estimate a dynamic panel

14Note that these definitions allow for banks to move between size categories over time.

9

regression:

4


4

4


j=0

j=0

+

+

4


4


j=0

3


k=1

4


j=0

17


k=1

∆log(List) = c +

αj ∆log(List−j) +

µj MPt−j + γ1INT RAst + γ2INT E Rst

j=1
j=0
φ j(MPt−j ∗ INT RAst) +

ϕj(MPt−j ∗ INT E Rst) +

βj N AT ION ALt−j

δj ST AT Est−j +

ψkQU ART E Rkt +

ξkY E ARkt + ηi + 
ist

(1.2)
where the dependent variable is real loan growth of bank i, located in state s, in quarter t. The
independent variables include 4 lags of bank i’s loan growth, the contemporaneous value and 4
lags of monetary policy shocks, a dummy variable equaling 1 if state s permits in-state branching
via M&A in quarter t, an analogous dummy variable equaling 1 if interstate banking is allowed
in state s during quarter t, and interactions between the monetary policy shocks and deregulation
dummies. Also included are the contemporaneous values and 4 lags of national and state control
variables, quarter dummy variables, year dummy variables, and a bank fixed effect.

The national-level variables include change in real GDP, change in the personal consumption
expenditures (PCE) index, and the CRSP value weighted stock return index. The state-level
variables include percentage change in personal income and change in the U.S. Federal Housing
Finance Agency all-transactions house price index. Quarter dummies are included to control for
seasonality in lending. Year dummies are included to control for additional macro-level phenomena
occurring during this time period, e.g.
the gradual phaseout of regulation Q, the Fed regime of
targeting non-borrowed reserves, and the Great Moderation. Alternate specifications with varying
levels of fixed effects and time dummies are presented below, with the most comprehensive dropping
all national-level variables in favor of quarterly fixed effects.

The coefficients of interest are the sum of the φ(cid:48)

intrastate branching deregulation. A significant
4

j=0 φ j
would indicate that monetary policy has a significantly different impact on bank lending following
j=0 ϕj would indicate the same for interstate

j s and sum of the ϕ(cid:48)

j s. A significant
4

10

banking deregulation. We have no prior expectation regarding the sign of the coefficients, as the
effect of deregulation on loan sensitivity to monetary policy is theoretically ambiguous.

1.4 Results

1.4.1

Impact of Deregulation

Equation 1.2 is estimated over the sample 1976Q2 - 1994Q4. Results for the summed coefficients
of interest are presented in panel (a) of table 1.4. Results for all coefficients are presented in table
1.5. Columns (1) through (5) in table 1.4 display results from a variety of specifications, with
column (4) reporting the baseline specification depicted in equation 1.2. Columns (1)-(3) provide
results for more loosely specified variations of equation 1.2 and column (5) provides results for a
more tightly specified variation. In the first four columns the summed coefficients of the monetary
policy indicator are negative and jointly significant at the 1% level 15. A contractionary 100 basis
point exogenous monetary policy shock reduces lending by roughly 1-2% over the following four
quarters 16. The summed coefficients on the interaction between intrastate branching deregulation
and monetary policy are small and insignificant in all five columns, indicating that intrastate
deregulation has no effect on loan sensitivity to monetary policy.

The summed coefficients on the interaction between the interstate banking deregulation dummy
and the monetary policy indicator are negative and significant in all five columns. An exogenous,
contractionary monetary policy shock reduces lending by an additional 1.38-4.26% for a bank
located in a state that has removed interstate banking restrictions 17. According to the first four
columns, the total effect of a contractionary monetary policy shock on lending for a bank located in
a deregulated state is a decline of 2.5-4.1%. The baseline specification in column (4) indicates that
the sensitivity of lending to monetary policy essentially doubles following interstate deregulation.
15Column (5) reports results including time fixed effects which are perfectly collinear with

national-level variables such as the monetary policy indicator.

16One standard deviation of the monetary policy indicator is 70 basis points, hence a contrac-

tionary one standard deviation shock reduces lending by 0.8-1.4% over the next four quarters.

17A contractionary one standard deviation shock reduces lending by an additional 1-3%.

11

Column (5) includes the strongest controls for time-specific macro variation, and indicates that the
effect of interstate deregulation is even stronger than that reported in column (4).

There is significant overlap in years that both types of restrictions are deregulated for a given
state. To check that inclusion of both sets of deregulation dummies is not biasing the results in
panel (a) equation 1.2 is estimated separately for each type of deregulation. Panel (b) presents
the summed coefficients of interest for estimating equation 1.2 with interstate deregulation dummy
and interactions only. Similarly, panel (c) presents results for estimating equation 1.2 with in-
trastate deregulation dummy and interactions only. Both panels are consistent with the baseline
results, confirming that lending becomes more sensitive to monetary policy after interstate banking
deregulation and that intrastate branching deregulation has no effect.

As discussed in section 1.3.1, our preferred measure of monetary policy is the RR shock series
rather than a traditional measure such as the fed funds rate. To check whether the baseline results are
driven by the choice of monetary policy indicator, equation 1.2 is also estimated with the quarterly
change in the fed funds rate as the monetary policy indicator. Results using the fed funds rate are
presented in panel (d). Once again, a contractionary monetary policy shock leads to a significant
decline in lending over the following four quarters. According to columns (1)-(4), a 100 basis point
increase in the federal funds rate leads to a 0.37-0.92% decline in lending over the next year.18
Intrastate deregulation once again has no effect.

Columns (1)-(3) of panel (d) report that lending is less sensitive to monetary policy after
interstate deregulation. Columns (4) and (5), which more completely control for unobserved macro
variation, are consistent with the results in panel (a) however. The summed coefficients in columns
(4) and (5) suggest that lending declines by an additional 0.66-1.02% after a state has removed
interstate restrictions. The smaller magnitudes and positive coefficients in columns (1)-(3) are
not surprising. The endogeneity and anticipatory components of the fed funds rate, which the
RR shocks control for, would naturally lead to a less pronounced effect or even opposite effect
of policy. Regardless, the richer specifications in panel (d) indicate that lending responds more
18This is in line with estimates from Ashcraft (2006), who finds that a 100 basis point increase

in the federal funds rate decreases bank lending by 0.45%.

12

strongly to policy after interstate deregulation, suggesting that choice of monetary policy variable
is not driving our results.

An additional concern raised in section 1.3.1 is that outliers in the monetary policy indicator
(as well as real loan growth) during the Fed’s period of non-borrowed reserve (NBR) targeting are
driving the results. To explicitly control for the NBR targeting period we estimate two other varia-
tions of equation 1.2 with results reported in table 1.6. Column (1) shows results including a NBR
dummy variable which equals 1 from 1979Q4-1982Q3 and 0 otherwise. Column (2) interacts the
NBR dummy with the contemporaneous value and lags of the monetary policy indicator. Column
(1) shows that the baseline results hold up when including the NBR dummy: lending declines by
1.71% prior to interstate deregulation and by an additional 2.33% after deregulation. The negative
and significant coefficient on the NBR dummy indicates that lending growth was lower during the
NBR targeting regime. Interestingly, in column (2) the summed coefficients on the monetary policy
indicator are positive and significant while the summed coefficients on the interaction between the
NBR dummy and monetary policy are negative and significant. Monetary policy therefore has a
strongly negative effect on lending during the NBR period. For our purposes, the key result is the
negative and significant coefficient on the interaction between interstate deregulation and monetary
policy, which confirms that the greater sensitivity of lending to policy after interstate deregulation
is not being driven by the NBR targeting regime.

The results in table 1.4 indicate that lending becomes more responsive to monetary policy
along the quantity dimension following interstate deregulation. Next, we examine how deregulation
impacts the sensitivity of lending to monetary policy along the price dimension. As discussed in
section 1.3.2, direct data on loan rates is not available through the Call Reports. We can proxy for
the average rate on a banks loan portfolio through the ratio of interest income on loans to quantity
of total loans however. This ratio is referred to as a bank’s average loan rate in the following.

Table 1.7 presents results for estimating equation 1.2 with average loan rate as the dependent
variable.19 Column (1) of table 1.7 shows that average loan rates significantly increase following a
19In the following we focus on the richer specifications including year dummies or time fixed

13

monetary tightening. For the four quarters following a 100 basis point exogenous and contractionary
monetary policy shock average loan rates increase by 69 basis points. The interaction between
the intrastate deregulation dummy and monetary policy is small and insignificant in both columns
(1) and (2), indicating that the removal of intrastate branching restrictions had no effect on the
sensitivity of loan pricing to monetary policy.

The interaction between the interstate deregulation dummy and monetary policy is positive and
significant in both columns, indicating that loan pricing becomes more sensitive to monetary policy
after interstate banking restrictions are removed. According to column (1), a bank located in a state
that has removed interstate restrictions increases its average loan rate by an additional 113 basis
points following a 100 basis point exogenous monetary tightening, which is more than double the
average increase for a bank in a state that has not deregulated. Column (2) reports a somewhat
smaller magnitude, indicating that a bank in a deregulated state increases its average loan rate by
an additional 47 basis points following a monetary tightening. Regardless, this is a meaningful
response as it is roughly two-thirds larger than that of a bank in a state which prohibits interstate
banking.

From 1976-1982 interest and fee income on loans is only reported in the second and fourth
quarters. For the above results we fill in the missing first and third quarter values so that first quarter
average loan rate is equal to the second quarter observation, and so that third quarter average loan
rate is equal to the fourth quarter observation. To check that replacing these missing values is not
driving the above results we re-estimate equation 1.2 for an abbreviated sample from 1983-1994.
Results are presented in columns (3) and (4) of table 1.7. The summed coefficients on the monetary
policy indicator in column (3) are no longer significant at the 10% level, but the magnitude is
similar and the standard errors are not large. The summed coefficients on the interaction between
the interstate deregulation dummy and monetary policy remain positive and significant in both
columns (3) and (4). These subsample results confirm that replacing the missing observations from
1976-82 is not driving the results in columns (1) and (2).
effects, i.e. the specifications corresponding to columns (4) and (5) in table 1.4.

14

1.4.2 Role of Bank Lending Channel

These results raise the question: why does interstate banking deregulation strengthen the effect of
monetary policy on lending? To investigate, we examine bank-level heterogeneity across a variety
of dimensions. The literature has established multiple characteristics which influence the strength
of the bank lending channel of monetary policy. Such characteristics include bank size (Kashyap
and Stein (1995)), liquidity (Kashyap and Stein (2000)), and capitalization (Kishan and Opiela
(2000)).
In this section we study the role of bank-level heterogeneity in explaining the greater
sensitivity of lending to monetary policy after interstate deregulation and the implications for the
lending channel of transmission.

Kashyap and Stein (1995) find that small banks are more sensitive to monetary policy than
larger banks. Similarly, Kashyap and Stein (2000) find that small and relatively illiquid banks are
most strongly affected by monetary policy. Cetorelli and Goldberg (2012) also report that global
banks are less responsive to monetary policy. We therefore investigate how heterogeneity across
size and domestic/foreign status is related to interstate deregulation’s impact on monetary policy
effectiveness. To investigate, we estimate equation 1.2 separately for small banks, large banks, and
branches of foreign banks operating in the United States. Consistent with the literature, we define
a small bank as any bank below the cross-sectional 95th percentile in total assets within a given
quarter. Correspondingly, a large bank is defined as any above the 95th percentile in total assets for
a given quarter.

Results are presented in panel (a) of table 1.8. Columns (1)-(2) have results for small banks,
(3)-(4) for large banks, and (5)-(6) for foreign banks. The summed coefficients in the first row
show that both small and large banks have a roughly 2% decline in lending for the four quarters
following a 100 basis point contractionary monetary policy shock prior to deregulation. The
response of lending from foreign banks is insignificant. The second row shows that intrastate
branching deregulation has no effect of the sensitivity of small bank lending, although it may
have a small negative effect on large bank lending. Interestingly, intrastate branching deregulation
seems to make branches of foreign banks much more sensitive to policy. The third row shows that

15

interstate banking deregulation only affects small banks. The coefficients are very similar to the
baseline results for all banks, as the response of small bank lending to a monetary shock doubles
after deregulation.

The results in panel (a) are consistent with Kashyap and Stein (1995) and Kashyap and Stein
(2000). As an additional check, we estimate 1.2 with average loan rate as the dependent variable for
both small and large bank samples. Foreign banks do not report interest and fee income on loans,
hence we cannot calculate average loan rates for them. Results are presented in table 1.9. The first
row of table 1.9 confirms that both small and large banks increase loan rates following a monetary
tightening. Columns (1) and (3) suggest that following interstate deregulation the sensitivity of
loan pricing to policy increases for both small and large banks. Column (2) confirms this for small
banks, however the summed coefficient for the interaction of interstate deregulation and monetary
policy is small and insignificant for large banks in column (4). These results provide further support
that the effect of interstate deregulation impacts monetary transmission through small banks.

Kashyap and Stein (2000) find that the bank lending channel operates through small and
relatively illiquid banks. To investigate the role of liquidity we estimate equation 1.2 by liquidity
quartile, where the 1st quartile includes the least liquid banks in a given quarter and the 4th quartile
includes the most liquid.20 Results are presented in table 1.10. Panel (a) displays results for all
banks. Prior to deregulation all four quartiles respond similarly to monetary policy, declining by
roughly 2% for the four quarters following a contractionary shock. After interstate deregulation
the least liquid banks are more strongly affected, with the most liquid banks in the 4th quartile not
responding any more strongly. Panel (b) presents results for small banks only and panel (c) for
large banks only. These panels confirm that the overall effect is being driven by the small banks.
According to the specification using time fixed effects all four quartiles become more sensitive to
policy after interstate deregulation, but the increase in response to policy is decreasing in liquidity.
Column (1) in panel (c) indicates the least liquid large banks may become slightly more sensitive
to policy after interstate deregulation, however column (2) does not confirm this. Overall, there is
20We use a narrower measure of liquidity - the ratio of cash and reserves to total liabilities - than

Kashyap and Stein (2000) for ease of interpretation.

16

little to no effect on large banks across liquidity quartiles.

Kishan and Opiela (2000) find that the effect of monetary policy on lending is stronger for
relatively undercapitalized banks, particularly small ones. We once again estimate equation 1.2,
this time by equity ratio quartile, where the 1st quartile includes the least capitalized banks in
a given quarter and the 4th quartile includes the most highly capitalized banks. Equity ratio is
calculated as total equity divided by total assets. Results are presented in table 1.11. Panel (a)
displays results for all banks. Prior to deregulation all four quartiles respond similarly to policy.
After interstate deregulation banks in all four quartiles become more responsive to policy, with an
additional decline in lending growth of 3.54-4.14% according to the specifications with time fixed
effects. The increased response is slightly larger for banks in the 1st and 2nd quartiles, but not large
enough enough to suggest that capitalization plays a major role in explaining the greater sensitivity
of lending to policy after interstate deregulation. Panel (b) presents results for small banks only
and panel (c) presents results for large banks. Once again, it is primarily small banks that become
more sensitive after deregulation.

1.5 Potential Explanations

The previous section establishes that following the removal of interstate banking restrictions
monetary policy has a stronger effect on lending for small and relatively illiquid banks. This implies
that interstate banking deregulation strengthens the bank lending channel of monetary transmission.
In this section we investigate three potential explanations for the strengthening of the lending channel
after deregulation. These explanations include changes in bank market structure, changes in loan
portfolio composition, and changes in bank-borrower relationships following deregulation.

1.5.1 Bank Market Structure

The effects of geographic banking deregulation have been widely discussed in the literature. Ja-
yaratne and Strahan (1998), Evanoff and Ors (2008) and Chortareas, Kapetanios, and Ventouri
(2016) find that deregulation increased efficiency in the banking sector. Stiroh and Strahan (2003)

17

report that deregulation improved competitive dynamics by reallocating market share to better per-
forming banks. Zou, Miller, and Malamud (2011) offer mixed evidence, arguing that deregulation
increased efficiency of small banks but decreased efficiency of medium-sized banks. Similarly,
Berger and DeYoung (2001) find both positive and negative links between geographic expansion
and bank efficiency. Rhoades (2000) argues that nationwide banking concentration increased from
1980-1998, in part due to geographic deregulation, and Jeon and Miller (2003) find that deregulation
is significantly correlated with higher state-level concentration.

Additionally, a relatively new literature has examined the relationship between banking market
structure and monetary policy transmission. In cross-country studies using bank-level measures
of market power Fungáčová, Solanko, and Weill (2014) and Leroy (2014) find that lending is
less sensitive to monetary policy when banks have greater market power. On the other hand,
Amidu and Wolfe (2013) and Yang and Shao (2016) find that lending is more sensitive to policy
when banks have greater market power. The only published study on bank market structure and
monetary transmission in the U.S. is Adams and Amel (2011) which takes market concentration at
the local level (MSA or county) as the measure of market structure. They use annual Community
Reinvestment Act data on new loan origination for a sample running from 1996-2004, and their
results show that monetary policy has a weaker effect on bank lending in more highly concentrated
markets.

The effect of geographic deregulation on banking market structure is not clear, nor is the effect
of market structure on monetary policy effectiveness.
In this section we test how each type of
deregulation impacted bank market power and banking concentration, at both the local and state
level. Additionally, we examine whether changes in market power and concentration can explain
the increased sensitivity of lending to policy following interstate deregulation. The measure of
market power used is a bank-level Lerner index and the measures of concentration used are the
Herfindahl-Hirschman Index (HHI), calculated at the local (county) and state levels.

The Lerner index is a measure of a banks market power, calculated as the difference between
price of output and marginal cost, divided by marginal cost. In calculating the Lerner index we

18

follow Fungáčová, Solanko, and Weill (2014) among others. The average price of bank production
is proxied by the ratio of total revenues to total assets. The marginal cost is calculated by estimating
a translog cost function with one output and three input prices. The output price is total assets and
the input prices are the price of labor, price of fixed assets, and price of borrowed funds (interest
on deposits). The cost function is specified as follows:

3

3


j=1

3


j=1

3


3


j=1

log(TCit) = α0 + α1log(yit) + 0.5α2(log(yit))2 +

βjlog(wj,it)

j=1
γjlog(yit) ∗ log(wj,it) + ρt + ηi + 
it

(1.3)

βjklog(wj,it) ∗ log(wk,it) +

+

Where y is total assets and
3

k=1

j=1 wj are the three input prices. Quarter dummies and bank fixed
effects are included. Symmetry and linear homogeneity restrictions are imposed on input prices.
Total cost is the sum of the three input prices. Marginal cost can then be calculated from the
estimated coefficients:

MC = (TC/y) ∗ (α1 + α2log(y) +

log(wj))

(1.4)

The resulting Lerner index, calculated as (P-MC)/MC, is a bank-level measure of market power,
with a value of 0 representing a perfectly competitive bank (P=MC) and a value of 1 representing
a pure monopolist. Since expense data is available only biannually until 1983 we fill the missing
first and third quarter observations with the average Lerner Index of the previous and following
quarters.

The Herfindahl-Hirschman Index (HHI) is calculated as the summed squares of firm market

shares within an industry:

HHI =

where s is the market share of firm i and there are N banks in the market. Hence in a monopoly,
where a single banks’s market share is equal to 100%, the HHI index would be 1. On the opposite

2

si

(1.5)

N


i=1

19

end of the spectrum, the HHI for a decentralized market with many firms would be close to zero.
We calculate HHI at both the county and state levels, as concentration at the local level and at the
state level may be quite different.

Each of the three bank structure measures are regressed on the deregulation dummies and

controls in the following specification:

BMSist = c + γ1INT RAst + γ2INT E Rst + δST AT Est + βBANKist−1 + ρt + ηi + 
ist

(1.6)

Results are presented in panel (a) of table 1.12. The first row shows that intrastate branching
deregulation increased county-level banking concentration but had no effect on the Lerner index
or state-level concentartion 21. Interstate deregulation also increased county-level concentration,
along with increasing the Lerner index and decreasing state-level concentration. These results
indicate that the removal of interstate banking restrictions increased bank market power, increased
local concentration, and decreased state concentration.

Now that we have documented the effect of interstate deregulation on a variety of bank market
structure measures, we examine how these measures are related to the the sensitivity of loan pricing
to monetary policy. To do so we estimate an alternative version of equation 1.2, with the bank
market structure variables interacted with the monetary policy indicator.

∆log(List) = c +

αj ∆log(List−j) +

µj MPt−j + γBMSist +

φ j(MPt−j ∗ BMSist)


4

(1.7)
The summed coefficients on the interaction between bank market structure and monetary policy,
j=0 φ j, informs us of the differential response of bank loan pricing to monetary policy depending
21Note: the state HHI regression is ran at the state-level rather than the bank-level.

4


j=0

3


k=1

δj ST AT Est−j +

ψkQU ART E Rkt

4


j=1

4

17


j=0

k=1

+

+

βj N AT ION ALt−j +

ξkY E ARkt + ηi + 
ist

4

4


j=0

j=0

20

on a banks market power, local market concentration, and state concentration 22. Results are
presented in panel (b) of table 1.12. Columns (1), (3), and (5) show results for estimating equation
1.7 with each bank market structure variable without the deregulation dummy and interactions.
Columns (2),(4), and (6) show results for estimating equation 1.7 with each bank market structure
variable as well as the interstate deregulation dummy and interactions.

Across all six columns a contractionary monetary policy shock results in a decrease in lending
over the following four quarters, with the decrease being significant for all columns. The first
column reports that banks with a higher Lerner index (i.e. greater market power) are less sensitive
to monetary policy. A bank which is a pure monopolist (Lerner = 1) decreases lending by 0.48%
for the four quarters following a monetary tightening whereas a perfectly competitive bank (Lerner
= 0) decreases lending by 2.91%. Column (2) confirms that the effect of policy on lending
increases by roughly 2% after interstate deregulation as in the baseline results. Columns (3)-(6)
report that county-level and state-level concentration have no effect on loan response to monetary
policy. Column (4) and column (6) also confirm that interstate banking deregulation increases the
sensitivity of lending to monetary policy.

The results in table 1.12 indicate that interstate deregulation did not affect loan sensitivity
through banking competition or market structure. Banking concentration has no impact on the
sensitivity of lending to monetary policy. Increased bank market power weakens the impact of
policy on lending. Since interstate deregulation increased bank market power but strengthened the
impact of policy on lending, the effect of deregulation could not have been driven by change in
market power. As a final investigation we estimate equation 1.2 for subsamples corresponding to
Lerner index quartile and local HHI quartile. While deregulation may not have operated through
increasing market power or concentration it is possible that banks were asymmetrically impacted
depending on their competitive environment. We therefore investigate the role of market structure
heterogeneity on the effect of deregulation.

Table 1.13 shows the effect of deregulation by Lerner index quartile, with the 1st quartile having
22Note: The Lerner Index is included with one lag to reduce simultaneity concerns.

21

the lowest market power (and hence being relatively more competitive) and with the 4th quartile
having the highest market power (and hence being relatively less competitive). Panel (a) shows
results for all banks, panel (b) shows results for small banks only, and panel (c) shows results for
large banks only. There is no clear trend across quartiles, as all four respond more strongly after
interstate deregulation, particularly the first and fourth quartiles. Once again, only small banks
respond more strongly after interstate deregulation, as there is no effect for large banks. Similarly,
table 1.14 shows the effect of deregulation by county HHI, with the 1st quartile having the lowest
concentration and the 4th quartile having the highest concentration. Panel (a) shows results for
all banks, panel (b) shows results for small banks, and panel (c) shows results for large banks.
Interstate deregulation has a significant effect across all four quartiles, once again driven by small
banks. Bank market structure therefore seems to play no role in the greater sensitivity of lending
to monetary policy after the removal of interstate restrictions.

1.5.2 Loan Portfolio Composition

Den Haan, Sumner, and Yamashiro (2007) find that certain types of loans are more sensitive
to monetary policy than others.
Interstate deregulation may therefore increase certain types of
lending which are more sensitive to policy. Den Haan, Sumner, and Yamashiro (2007) examine
loan portfolio response to monetary policy at the aggregate level and find differential responses
depending on loan type. Real estate and consumer loans decrease following a monetary tightening
but commercial and industrial (C&I) loans actually increase. In explaining these results the authors
suggest that adjusting loan portfolio composition may be an optimal response to monetary shocks
for a variety of reasons.

Focusing on total lending may therefore hide important compositional effects. First, we check
whether interstate deregulation altered the average composition of a bank’s loan portfolio. Panel (a)
of table 1.15 shows the effect of deregulation on the average share of each loan category (relative
to total loans).
Intrastate and interstate deregulation both significantly decrease C&I and real
estate lending as a share of total loans. Interstate deregulation significantly increases the share of

22

consumer lending. The coefficients for each category share are small however, as share of loans
going to consumer lending increases by just 0.37% after deregulation. This makes it implausible
that a change in loan portfolio composition is driving the baseline results.

To investigate further, equation 1.2 is estimated separately for each of the three main loan
categories (C&I, real estate, and consumer) with results presented in panel (b) of table 1.15.
Interestingly, and inconsistent with Den Haan, Sumner, and Yamashiro (2007), columns (1), (3),
and (5) report that each loan category responds negatively to a monetary tightening 23. While
the summed coefficients on the interaction between interstate deregulation and monetary policy
are not significant for the baseline specification, the alternate specification including time fixed
effects shows that each category becomes more sensitive to policy following deregulation, and at
a similar magnitude as total lending in table 1.4. Since each type of lending responds to interstate
deregulation in a similar manner it appears that the greater sensitivity of overall lending to monetary
policy after deregulation cannot be explained by changes in loan portfolio composition.

1.5.3 Bank-Borrower Relationships

A third potential explanation for our results is a dilution of bank-borrower customer relationships
and a greater propensity of banks to cut lending following interstate deregulation. Petersen and
Rajan (1994) find that bank-borrower relationships increase credit availability for bank customers.
Cole, Goldberg, and White (2004) find that large banks make lending decisions based on standard
financial criteria whereas small banks employ greater discretion based on impressions of borrower
characteristics. Bonaccorsi di Patti and Gobbi (2001) report that acquisition and entry into new
markets tends to reduce credit supply. These results suggest that increased entry into new markets
and increased concentration following deregulation may weaken relationships between banks and
borrowers, leading to a greater decline in lending following a contractionary monetary shock.

To investigate we estimate equation 1.2 for two subsamples: banks that are affiliated with a
23There are important differences between this study and Den Haan, Sumner, and Yamashiro
(2007) however as they use aggregate data in a VAR framework for a sample that extends to 2004.

23

BHC and stand alone banks that are unaffiliated with a holding company. Affiliated banks are
more likely to operate within a centralized organizational structure that is less sensitive to local
borrower characteristics. If the strengthening of the bank lending channel is driven by a weakening
of bank-borrower relationships we would therefore expect affiliated banks to be more strongly
affected. Results are presented in panel (a) of table 1.16. Columns (1)-(2) show results for stand
alone banks and columns (3)-(4) show results for BHC affiliated banks. Ashcraft (2006) has
previously found that the bank lending channel is stronger for stand alone banks than for affiliated
banks. Consistent with these results, columns (1) and (3) show that stand alone banks respond
more strongly to monetary policy than affiliated banks pre-deregulation. However after interstate
deregulation affiliated banks become significantly more sensitive than stand alone banks. Column
(1) indicates that interstate deregulation does not significantly impact stand alone banks. Column (2)
suggests stand alone banks become somewhat more sensitive to policy after deregulation. Columns
(3) and (4) indicate that affiliated banks become significantly more responsive to monetary policy
after deregulation, by a relatively large magnitude of 2.7-5.16%. Panel (b) presents results for small
banks only and panel (c) for large banks only. Once again, the overall effect is driven by small
banks, as the BHC affiliated small banks respond more strongly than stand alone small banks.

Heterogeneity across bank size, liquidity, and BHC affiliation therefore appear to be driving
the effect of interstate deregulation on monetary policy. Small banks, less liquid banks, and banks
affiliated with a BHC are most strongly impacted by deregulation. Therefore, we further split
the sample to account for all three characteristics. Equation 1.2 is estimated across liquidity ratio
quartile for four groups: small affiliated banks, small stand alone banks, large affiliated banks,
and large stand alone banks. Panel (a) of table 1.17 shows the effect of interstate deregulation by
liquidity quartile for small BHC-affiliated banks; panel (b) for small stand alone banks; panel (c)
for large BHC-affilaited banks; and panel (d) for large stand alone banks. The results confirm that
deregulation primarily leads to small and affiliated banks becoming more sensitive to monetary
policy, and that the effect is decreasing in liquidity.24

24The least liquid small stand alone banks and the 2nd quartile of large stand alone banks also

24

Next, we test why small affiliated banks are most strongly affected. To do so we look more
broadly at the asset side of a bank’s balance sheet. We once again estimate equation 1.2 across
liquidity quartile for small banks based upon their BHC affiliation, but now with total asset growth
as the dependent variable in one specification and with securities growth as the dependent variable
in a second specification. Table 1.18 presents results. Panel (a) shows how interstate deregulation
impacts the sensitivity of asset growth to monetary policy for small affiliated banks. Panel (b)
shows the same effect for small stand alone banks. Panel (c) shows how interstate deregulation
impacts the sensitivity of securities growth to monetary policy for small affiliated banks. Panel (d)
shows the same effect for small stand alone banks. Noticeably, small affiliated banks adjust both
assets and securities in response to a monetary shock whereas small stand alone banks do not adjust
either. Small affiliated banks below the 4th quartile in liquidity see a relatively small decline in
assets for the four quarters following a monetary policy shock. As seen in table 1.17, these banks
are reducing lending to a relatively large degree. On the other hand, panel (c) in table 1.18 shows
that they increase securities holdings in response to a contractionary shock (except for the 3rd
quartile), resulting in an overall small decline in assets. Thus it appears that relatively illiquid small
banks affiliated with a BHC respond uniquely to monetary policy after interstate deregulation, by
shifting the asset-side of their balance sheets away from lending and towards securities.

We find that the bank lending channel of transmission strengthens primarily for small, relatively
illiquid banks affiliated with a BHC. These banks are unique in more strongly adjusting the asset-
side of their balance sheets towards securities and away from loans following a contractionary
monetary shock. These results are consistent with the notion that interstate deregulation weakens
bank-borrower relationships and increases the propensity of banks to cut lending in response to a
contractionary monetary shock.
become more sensitive to policy after deregulation. There are very few large stand alone banks,
hence the relatively large and weakly significant coefficient for the 2nd quartile in panel (d) should
be interpreted with caution.

25

1.6 Aggregate Effects

We have documented that bank lending becomes more sensitive to monetary policy following
interstate banking deregulation, particularly for small banks and banks affiliated with a bank holding
company. In this section we aggregate our bank-level data to the state-level to investigate the state-
level effects of deregulation. Table 1.20 presents results for estimating equation 1.2 with state-level
variables.25 Column (1) presents results with state-level real loan growth from all banks as the
dependent variable. The interaction between interstate deregulation and monetary policy is negative
but insignificant, indicating that deregulation has no effect on aggregate state-level loan growth.

To investigate further we aggregate state loans separately for four different categories of banks.
The four categories are the same as those focused on in section 1.5.3: small BHC affiliated banks,
small stand alone banks, large BHC affiliated banks, and large stand alone banks. Table 1.19
presents summary statistics for the share of total loans from each type of bank. Figure 1.5 plots
each groups loan share over the entire sample. For the whole sample 61% of total lending comes
from large affiliated banks with small affiliated banks having the next largest share at 16%. The
average share of small stand alone banks drops by half, from 16% to 8% from the early part of the
sample to the later part. The shares of the other three groups increase over this time frame, with
small affiliated banks having the largest increase from 13% to 18%.

We once again estimate equation 1.2 with aggregate loan growth from each of the four bank
categories as the dependent variable. Results are presented in table 1.20. Interstate deregulation
only impacts aggregate lending from small affiliated banks. Following a 100 basis point contrac-
tionary shock lending growth from all small and affiliated banks within a state declines by 8% over
the following four quarters. Small affiliated banks make up on average 16% of total lending over
the sample, hence a back-of-the-envelope calculation indicates that after interstate deregulation
state lending growth declines by an additional (8% x 0.16) = 1.28% following a contractionary
shock. This rough estimate is in-line with the summed coefficients on the interstate-monetary
policy interaction term in the first column of table 1.20 for all banks. While the magnitude is not
25The results shown are from the specification with time fixed effects rather than year dummies.

26

large, it is noteworthy that interstate banking deregulation results in a greater response of state-level
loan growth to monetary policy in addition to the stronger response at the individual bank-level.

1.7 Conclusion

This paper examines the relationship between bank regulation and monetary policy. From the
mid-1970’s to mid-1990’s a majority of states removed restrictions on geographic bank expansion.
There were two types of restrictions: those on out-of-state ownership of in-state banks (interstate)
and those on within-state branching (intrastate). By exploiting the staggered timing of state-level
deregulation we find that interstate banking deregulation, but not intrastate branching deregulation,
increases the sensitivity of lending to monetary policy. The response of real loan growth to
monetary policy doubles following interstate deregulation.

More specifically, interstate banking deregulation strengthens the bank lending channel of
monetary transmission, as monetary policy has a greater effect on small and relatively illiquid banks
after deregulation. We consider a variety of explanations for these results. Though deregulation
increases bank market power and local banking concentration, neither of these changes in bank
market structure can explain the increased sensitivity of loans to monetary policy. Deregulation
impacts all three major loan categories similarly, also ruling out the possibility that the greater
response of lending is driven by changes in loan portfolio composition.

On the other hand, we find that banks affiliated with a bank holding company are most strongly
impacted by deregulation. After deregulation small banks affiliated with a bank holding company
respond to monetary policy through a larger substitution of securities for bank loans. These
results point to a weakening of bank-borrower relationships and an increased propensity of banks
to cut lending as the mechanism behind interstate deregulation’s strengthening of monetary policy
transmission. Finally, we find that interstate banking deregulation leads to a greater effect of
monetary policy on loan growth at the aggregate state-level in addition to at the individual bank-
level. Further investigation into the aggregate effects of interstate banking deregulation remains an
intriguing avenue for future work.

27

APPENDIX

28

Table 1.1: Timing of State Deregulation

State
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Washington, DC
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming

Intrastate branching via M&A

Interstate banking

1981
<1970
<1970
1994
<1970
1991
1980
<1970
<1970
1988
1983
1986
<1970
1988
1989

*

1987
1990
1988
1975
<1970
1984
1987
1993
1986
1990
1990
1985
<1970
1987
1977
1991
1976
<1970
1987
1979
1988
1985
1982
<1970
<1970
<1970
1985
1988
1981
1970
1978
1985
1987
1990
1988

1987
1982
1986
1989
1987
1988
1983
1988
1985
1985
1985

*

1985
1986
1986
1991
1992
1984
1987
1978
1985
1983
1986
1986
1988
1986
1993
1990
1985
1987
1986
1989
1982
1985
1991
1985
1987
1986
1986
1984
1986
1988
1985
1987
1984
1988
1985
1987
1988
1987
1987

Column 1 lists the year that each state allowed branch banking through mergers and acquisitions.
Column 2 lists the year each state entered into an interstate banking agreement with other states. *
indicates that a state had not deregulated before 1994.

29

Table 1.2: Summary Statistics - All Banks

All banks

Whole Sample
Mean

Std Dev

1976-1985

1986-1994

Mean

Std Dev

Mean

Std Dev

Real loan growth - Total (%)
Avg loan rate (annualized %)

1.13
11.43

Real loan growth - C&I (%)
Real loan growth - RE (%)
Real loan growth - Con (%)
C&I Share of Lending
RE Share of Lending
Con share of lending

0.76
2.16
0.46
0.21
0.40
0.24

(7.25)
(4.06)

(24.13)
(15.08)
(17.35)
(0.14)
(0.19)
(0.14)

1.39
11.99

1.63
2.18
1.04
0.22
0.34
0.26

(7.30)
(4.77)

(24.32)
(16.77)
(17.58)
(0.14)
(0.17)
(0.14)

0.81
10.74

-0.33
2.13
-0.27
0.20
0.47
0.20

(7.17)
(2.83)

(23.85)
(12.66)
(17.03)
(0.13)
(0.18)
(0.13)

Assets ($)
Securities ($)
Liquidity Ratio
Equity Ratio
BHC Affiliation

Lerner Index
County HHI
State HHI

173 mil
38 mil
0.09
0.09
0.53

(2 bil)
(269 mil)
(0.23)
(0.03)
(0.50)

122 mil
27 mil
0.10
0.09
0.39

(1.6 bil)
(183 mil)
(0.06)
(0.03)
(0.49)

235 mil
54 mil
0.08
0.09
0.70

(2.5 bil)
(355 mil)
(0.34)
(0.04)
(0.46)

0.31
0.33
0.11

(0.09)
(0.22)
(0.11)

0.30
0.31
0.10

(0.08)
(0.21)
(0.10)

0.32
0.35
0.12

(0.09)
(0.23)
(0.12)

Number of banks

16,014

14,835

14,242

This table reports summary statistics for bank-level variables of interest. The first two columns have
statistics for the entire sample (1976Q2 - 1994Q4). The third and fourth columns have statistics for
the early part of the sample (when the majority of states had not deregulated). The fifth and sixth
columns have statistics for the later part of the sample (when the majority of states had deregulated).

30

Table 1.3: Summary Statistics - by Bank Size

Panel (a): small banks

Whole Sample
Mean

Std Dev

1976-1985

1986-1994

Mean

Std Dev

Mean

Std Dev

Real loan growth - Total (%)
Avg loan rate (annualized %)

1.15
11.43

Real loan growth - C&I (%)
Real loan growth - RE (%)
Real loan growth - Con (%)
C&I Share of Lending
RE Share of Lending
Con share of lending

Assets($)
Securities ($)
Liquidity Ratio
Equity Ratio
BHC Affiliation

Lerner Index
County HHI
State HHI

0.78
2.20
0.48
0.21
0.40
0.24

51 mil
17 mil
0.10
0.09
0.51

0.30
0.33
0.11

(7.31)
(0.04)

(24.51)
(15.30)
(17.54)
(0.14)
(0.19)
(0.14)

(58 mil)
(20 mil)
(0.24)
(0.03)
(0.50)

(0.09)
(0.22)
(0.11)

1.40
11.98

1.66
2.24
1.07
0.22
0.35
0.27

38 mil
13 mil
0.10
0.09
0.37

0.29
0.31
0.10

(7.38)
(4.77)

(24.79)
(17.09)
(17.85)
(0.14)
(0.17)
(0.14)

(40 mil)
(14 mil)
(0.06)
(0.03)
(0.48)

(0.08)
(0.21)
(0.10)

0.83
10.74

-0.32
2.16
-0.26
0.20
0.47
0.20

68 mil
23 mil
0.08
0.09
0.69

0.31
0.35
0.12

(7.20)
(2.63)

(24.12)
(12.70)
(17.11)
(0.13)
(0.18)
(0.13)

(72 mil)
(26 mil)
(0.35)
(0.04)
(0.46)

(0.09)
(0.23)
(0.12)

Number of banks
Panel (b): large banks

15,481

14,264

13,625

Mean

Std Dev

Mean

Std Dev

Mean

Std Dev

Real loan growth - Total (%)
Avg loan rate (annualized %)

0.86
11.53

Real loan growth - C&I (%)
Real loan growth - RE (%)
Real loan growth - Con (%)
C&I Share of Lending
RE Share of Lending
Con share of lending

0.39
1.30
0.08
0.30
0.36
0.24

(6.09)
(5.09)

(15.16)
(10.07)
(13.35)
(0.14)
(0.17)
(0.16)

1.13
12.17

1.15
1.09
0.44
0.32
0.32
0.25

(5.63)
(4.78)

(12.59)
(8.50)
(11.42)
(0.13)
(0.14)
(0.12)

0.52
10.74

-0.59
1.58
-0.38
0.27
0.41
0.23

(6.61)
(5.35)

(17.86)
(11.75)
(15.41)
(0.15)
(0.19)
(0.19)

Assets($)
Securities ($)
Liquidity Ratio
Equity Ratio
BHC Affiliation

Lerner Index
County HHI
State HHI

2.5 bil
447 mil
0.12
0.07
0.83

(8.7 bil)
(1.1 bil)
(0.08)
(0.02)
(0.37)

1.7 bil
306 mil
0.14
0.07
0.75

(6.8 bil)
(765 mil)
(0.08)
(0.02)
(0.43)

3.4 bil
640 mil
0.09
0.07
0.94

(10.5 bil)
(1.5 bil)
(0.07)
(0.03)
(0.24)

0.37
0.33
0.11

(0.11)
(0.20)
(0.11)

0.34
0.31
0.10

(0.09)
(0.19)
(0.10)

0.41
0.35
0.12

(0.12)
(0.21)
(0.12)

Number of banks

1,215

931

1,023

This table reports summary statistics for bank-level variables of interest. Panel (a) reports statistics
for small banks, defined as all banks under the 95th percentile in total assets in a given quarter.
Panel (b) reports statistics for large banks, defined as all banks above the 95th percentile in total
assets in a given quarter. The first two columns have statistics for the entire sample (1976Q2 -
1994Q4). The third and fourth columns have statistics for the early part of the sample (when the
majority of states had not deregulated). The fifth and sixth columns have statistics for the later part
of the sample (when the majority of states had deregulated).

31

Table 1.4: Impact of Deregulation on Lending Response to Monetary Policy

sum of coefficients
Panel (a): baseline results

(1)

(2)

(3)

(4)

MP

-0.0111***
(0.0014)

-0.0115***
(0.0013)

-0.0121***
(0.0014)

-0.0202***
(0.0021)

(5)

-

Intra*MP

Inter*MP

0.0023
(0.0029)

0.0026
(0.0027)

0.0022
(0.0029)

-0.0005
(0.0030)

-0.0010
(0.0031)

-0.0142***
(0.0046)

-0.0139***
(0.0043)

-0.0138***
(0.0045)

-0.0208**
(0.0094)

-0.0426***
(0.0112)

Panel (b): interstate deregulation only

MP

Inter*MP

-0.0109***
(0.0014)

-0.0126***
(0.0041)

-0.0109***
(0.0014)

-0.0122***
(0.0040)

-0.0116***
(0.0015)

-0.0124***
(0.0042)

-0.0203***
(0.0022)

-

-0.0209**
(0.0084)

-0.0424***
(0.0113)

Panel (c): intrastate deregulation only

MP

-0.0108***
(0.0013)

-0.0112***
(0.0012)

-0.0118***
(0.0012)

-0.0205***
(0.0018)

-

Intra*MP

-0.0013
(0.0025)

-0.0008
(0.0024)

-0.0012
(0.0026)

-0.0023
(0.0027)

-0.0032
(0.0033)

Panel (d): fed funds rate as MP indicator

MP

-0.0092***
(0.0007)

-0.0091***
(0.0008)

-0.0099***
(0.0009)

-0.0037***
(0.0014)

-

Intra*MP

Inter*MP

observations
STATE
NATIONAL
State Fixed Effects
Bank Fixed Effects
Linear Time Trend
Year Dummies
Time Fixed Effects

0.0015
(0.0011)

0.0014
(0.0012)

0.0017
(0.0012)

0.0006
(0.0010)

0.0006
(0.0010)

0.0051***
(0.0014)

0.0051***
(0.0014)

0.0053***
(0.0015)

-0.0066**
(0.0028)

-0.0102***
(0.0030)

823,659

823,659

823,659

823,659

823,659

Yes
Yes
-
-
Yes
-
-

Yes
Yes
Yes
-
Yes
-
-

Yes
Yes
-
Yes
Yes
-
-

Yes
Yes
-
Yes
-
Yes
-

Yes
-
-
Yes
-
-
Yes

This table reports results from estimating equation 1.2. Panel (a) reports the baseline results. Panel
(b) reports results for estimating equation 1.2 with interstate deregulation only and panel (c) reports
results for estimating equation 1.2 with intrastate deregulation only. Panel (d) reports results using
the quarterly change in the fed funds rate as the monetary policy indicator, rather than the RR
shocks. Robust standard errors clustered at the state-level are in parentheses. * indicates statistical
significance at the 10% level. ** indicates statistical significance at the 5% level. *** indicates
statistical significance at the 1% level.

32

Table 1.5: Impact on Lending Response: All Coefficients - Baseline Model

Variable

Loan Growth (t-1)

Loan Growth (t-2)

Loan Growth (t-3)

Loan Growth (t-4)

GDP

GDP(t-1)

GDP(t-2)

GDP(t-3)

GDP(t-4)

PCE

PCE(t-1)

PCE(t-2)

Coefficient
0.100***
(0.0126)

0.0157
(0.0105)

0.0395***
(0.00487)

0.166***
(0.0114)

-4.85e-07
(6.31e-06)

-1.64e-06
(8.31e-06)

-4.23e-05***
(6.47e-06)

3.73e-05***
(5.19e-06)

1.29e-05*
(6.66e-06)

-0.00388*
(0.00214)

-0.000401
(0.00229)

0.0114***
(0.00263)

PCE(t-3)

-0.0229***
(0.00295)

PCE(t-4)

CRSP

CRSP(t-1)

CRSP(t-2)

CRSP(t-3)

-0.00294
(0.00274)

0.0156
(0.0115)

0.0553***
(0.0156)

0.0552***
(0.0143)

0.102***
(0.0144)

Dependent variable: Real Loan Growth (1976Q2 - 1994Q4)
Coefficient
Variable
-0.000882
CRSP(t-4)
(0.000764)

Coefficient
0.0491**
(0.0186)

Variable
INTRA*MP

PI

0.000959***
(0.000293)

INTRA*MP(t-1)

PI(t-1)

PI(t-2)

PI(t-3)

0.00150***
(0.000403)

0.00187***
(0.000299)

0.000437*
(0.000220)

INTRA*MP(t-2)

INTRA*MP(t-3)

INTRA*MP(t-4)

-0.000437
(0.000745)

0.000659
(0.000844)

0.000435
(0.000844)

-0.000322
(0.000927)

PI(t-4)

0.000708***
(0.000209)

INTER*MP -0.00878***
(0.00262)

HPI

0.000477***
(8.85e-05)

INTER*MP(t-1)

HPI(t-1)

HPI(t-2)

HPI(t-3)

HPI(t-4)

0.000588***
(0.000107)

0.000726***
(8.75e-05)

0.000608***
(9.15e-05)

0.000333***
(6.50e-05)

MP

0.00224***
(0.000685)

MP(t-1)

MP(t-2)

MP(t-3)

MP(t-4)

INTRA

INTER

-0.00866***
(0.000613)

-0.00855***
(0.000600)

-0.00286***
(0.000456)

-0.00241***
(0.000440)

-0.000335
(0.00185)

0.00172
(0.00163)

INTER*MP(t-2)

INTER*MP(t-3)

INTER*MP(t-4)

Q2 Dummy

Q3 Dummy

Q4 Dummy

1978 Dummy

1979 Dummy

1980 Dummy

1981 Dummy

1982 Dummy

0.00108
(0.00261)

0.000507
(0.00243)

-0.00653***
(0.00213)

-0.00713***
(0.00197)

0.0208***
(0.00184)

0.00834***
(0.00150)

0.00400***
(0.00137)

-0.0120***
(0.00106)

-0.0283***
(0.00237)

-0.0296***
(0.00268)

0.0101**
(0.00420)

0.00766***
(0.00192)

Variable
1983 Dummy

1984 Dummy

1985 Dummy

1986 Dummy

1987 Dummy

1988 Dummy

1989 Dummy

Coefficient
0.00975***
(0.00267)

0.00646***
(0.00220)

-0.0129***
(0.00189)

-0.00734***
(0.00138)

-0.00888***
(0.00222)

0.00219
(0.00259)

-0.00105
(0.00245)

1990 Dummy

-0.00739***
(0.00237)

1991 Dummy

1992 Dummy

1993 Dummy

1994 Dummy

Constant

-0.00166
(0.00298)

-0.00758**
(0.00305)

-0.000698
(0.00392)

0.0147***
(0.00495)

-0.00467
(0.00510)

Observations

823,659

Number of banks
R-squared

15,990
0.124

This table reports full results from estimating equation 1.2 with the baseline specification. Robust
standard errors clustered at the state-level are in parentheses. * indicates statistical significance
at the 10% level. ** indicates statistical significance at the 5% level. *** indicates statistical
significance at the 1% level.

33

Table 1.6: Robustness: NBR Period

sum of coefficients

(1)

(2)

MP

-0.0171***
(0.0022)

0.0239***
(0.0079)

Intra*MP

Inter*MP

NBR

-0.0006
(0.0030)

-0.0233**
(0.0093)

-0.0048***
(0.0015)

NBR*MP

-

-0.0009
(0.0029)

-0.0340***
(0.0104)

-0.0125***
(0.0018)

-0.0406***
(0.0074)

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies

823,659

823,659

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

This table reports results from estimating equation 1.2 with controls for the period of non-borrowed
reserve (NBR) targeting from 1979-1982. Column (1) includes a dummy variable equalling 1 for
quarters during the NBR regime and equalling 0 otherwise. Column (2) includes the NBR dummy
and an interaction between the dummy and the monetary policy indicator. Robust standard errors
clustered at the state-level are in parentheses. * indicates statistical significance at the 10% level.
** indicates statistical significance at the 5% level. *** indicates statistical significance at the 1%
level.

34

Table 1.7: Impact on Loan Pricing Response to Monetary Policy

Dependent variable: Avg Loan Rate

sum of coefficients

(1)

(2)

1976-1994

1983-1994

(3)

(4)

MP

0.0069***
(0.0008)

-

0.0074
(0.0046)

-

Intra*MP

Inter*MP

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

0.0006
(0.0021)

0.0005
(0.0011)

0.0080*** 0.0081***
-0.0014
(0.0023)

0.0113***
(0.0020)

0.0047**
(0.0018)

0.0055*
(0.0028)

0.0036**
(0.0016)

822,792

822,792

494,975

494,975

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

This table reports results from estimating equation 1.2 with average loan rate as the dependent
variable. Columns (1) and (2) report results for the full sample with missing Q1 and Q3 ob-
servations filled for 1976-1982. Columns (3) and (4) report reports with an abbreviated sample
for robustness. Robust standard errors clustered at the state-level are in parentheses. * indicates
statistical significance at the 10% level. ** indicates statistical significance at the 5% level. ***
indicates statistical significance at the 1% level.

35

Table 1.8: Impact on Lending Response: by Bank Size

By bank size

sum of coefficients

Small Banks
(2)
(1)

Large Banks
(3)
(4)

Foreign Banks
(5)
(6)

MP

-0.0202***
(0.0022)

-

-0.0216***
(0.0037)

-

0.0157
(0.0644)

-

Intra*MP

Inter*MP

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

0.0001
(0.0033)

-0.0005
(0.0032)

-0.0212**
(0.0093)

-0.0439***
(0.0110)

787,027

787,027

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

-0.0056
(0.0034)

-0.0092
(0.0097)

36,632
Yes
Yes
Yes
Yes
-

-0.0058*
(0.0032)

-0.0081
(0.0131)

36,632
Yes
-
Yes
-
Yes

-0.0997***
(0.0231)

-0.0940***
(0.0252)

-0.0467
(0.0274)

-0.0417
(0.0390)

12,679
Yes
Yes
Yes
Yes
-

12,679
Yes
-
Yes
-
Yes

Panel (a) reports results from estimating equation 1.2 broken into three categories: small banks
(those under the 95th percentile in total assets), large banks (those above the 95th percentile in total
assets), and branches of foreign banks. Robust standard errors clustered at the state-level are in
parentheses. * indicates statistical significance at the 10% level. ** indicates statistical significance
at the 5% level. *** indicates statistical significance at the 1% level.

36

Table 1.9: Impact on Loan Pricing: by Bank Size

Dependent variable: Avg loan rate

sum of coefficients

(1)

Small

MP

0.0067***
(0.0009)

(2)

-

Large

(3)

0.0110***
(0.0014)

(4)

-

Intra*MP

Inter*MP

0.0010
(0.0022)

0.0009
(0.0012)

-0.0045*
(0.0022)

-0.0045***
(0.0010)

0.0111***
(0.0020)

0.0042**
(0.0019)

0.0129***
(0.0029)

0.0011
(0.0020)

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

786,207

786,207

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

36,585
Yes
Yes
Yes
Yes
-

36,585
Yes
-
Yes
-
Yes

This table reports results from estimating equation 1.2 with average loan rate as the dependent
variable. Columns (1) and (2) report results for small banks only (those under the 95th percentile
in total assets). Columns (3) and (4) report reports for large banks only (those above the 95th
percentile in total assets). Robust standard errors clustered at the state-level are in parentheses.
* indicates statistical significance at the 10% level. ** indicates statistical significance at the 5%
level. *** indicates statistical significance at the 1% level.

37

Table 1.10: Impact on Lending Response: by Bank Liquidity

All banks - by liquidity ratio quartile

Panel (a):
sum of coefficients

1st

(1)

MP

-0.0194***
(0.0027)

(2)

-

2nd

(3)

-0.0206***
(0.0027)

(4)

-

3rd

(5)

-0.0207***
(0.0022)

(6)

-

4th

(7)

-0.0219***
(0.0030)

(8)

-

Intra*MP

Inter*MP

0.0064
(0.0043)

0.0052
(0.0049)

0.0023
(0.0031)

0.0015
(0.0030)

-0.0332**
(0.0128)

-0.0625***
(0.0166)

-0.0200*
(0.0109)

-0.0491***
(0.0106)

-0.0006
(0.0028)

-0.0127
(0.0090)

-0.0007
(0.0023)

-0.0259**
(0.0104)

-0.0070
(0.0044)

-0.0084
(0.0100)

-0.0072**
(0.0036)

-0.0243
(0.0152)

observations

208,273

208,273

207,506

207,506

205,939

205,939

201,941

201,941

Small banks - by liquidity ratio quartile

Panel (b):
sum of coefficients

1st

(1)

MP

-0.0196***
(0.0027)

(2)

-

2nd

(3)

-0.0207***
(0.0027)

(4)

-

3rd

(5)

-0.0208***
(0.0022)

(6)

-

4th

(7)

-0.0215***
(0.0033)

(8)

-

Intra*MP

Inter*MP

0.0067
(0.0044)

0.0055
(0.0048)

0.0025
(0.0033)

0.0017
(0.0030)

-0.0340***
(0.0125)

-0.0635***
(0.0172)

-0.0200*
(0.0110)

-0.0502***
(0.0106)

0.0007
(0.0029)

-0.0121
(0.0094)

0.0005
(0.0024)

-0.0257**
(0.0107)

-0.00072
(0.0051)

-0.0095
(0.0104)

-0.0076*
(0.0038)

-0.0267*
(0.0151)

observations

204,559

204,559

201,126

201,126

195,307

195,307

186,035

186,035

Large banks - by liquidity ratio quartile

Panel (c):
sum of coefficients

MP

Intra*MP

Inter*MP

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

(1)

-0.0106
(0.0198)

0.015
(0.0181)

-0.0308
(0.0311)

3,714
Yes
Yes
Yes
Yes
-

1st

(2)

-

0.0115
(0.0168)

-0.0460
(0.0538)

3,714
Yes
-
Yes
-
Yes

2nd

(3)

-0.0212***
(0.0073)

0.0025
(0.0074)

-0.0169
(0.0234)

6,380
Yes
Yes
Yes
Yes
-

(4)

-

0.0033
(0.0072)

0.0138
(0.0228)

6,380
Yes
-
Yes
-
Yes

3rd

(5)

-0.0195***
(0.0053)

(6)

-

-0.0111*** -0.0109***
(0.0037)
(0.0034)

0.0018
(0.0120)

0.0173
(0.0112)

10,362
Yes
Yes
Yes
Yes
-

10,362
Yes
-
Yes
-
Yes

4th

(7)

-0.0261***
(0.0054)

(8)

-

-0.0044
(0.0059)

-0.0053
(0.0157)

15,906
Yes
Yes
Yes
Yes
-

-0.0048
(0.0041)

-0.0253
(0.0225)

15,906
Yes
-
Yes
-
Yes

Panel (a) reports results from separately estimating equation 1.2 for all banks that fall into the 1st,
2nd, 3rd, and 4th quartiles of liquidity ratio within a given quarter. Panel (b) reports results for
small banks only and panel (c) reports results for large banks only. Liquidity ratio is measured as
total cash and reserves divided by total liabilities. Robust standard errors clustered at the state-level
are in parentheses. * indicates statistical significance at the 10% level. ** indicates statistical
significance at the 5% level. *** indicates statistical significance at the 1% level.

38

Table 1.11: Impact on Lending Response: by Bank Capitalization

All banks - by equity ratio quartile

Panel (a)
sum of coefficients

1st

(1)

MP

-0.0206***
(0.0025)

(2)

-

2nd

(3)

-0.0190***
(0.0021)

(4)

-

3rd

(5)

-0.0211***
(0.0029)

(6)

-

4th

(7)

-0.0200***
(0.0035)

(8)

-

Intra*MP

Inter*MP

-0.0009
(0.0034)

-0.0009
(0.0028)

0.0015
(0.0035)

0.0008
(0.0034)

-0.0219*
(0.0113)

-0.0427***
(0.0109)

-0.0235**
(0.0104)

-0.0414***
(0.0108)

-0.0032
(0.0042)

-0.0134
(0.0099)

-0.0037
(0.0033)

-0.0354***
(0.0117)

0.0035
(0.0031)

-0.0096
(0.0109)

0.0026
(0.0039)

-0.0369**
(0.0165)

observations

204,740

204,740

208,100

208,100

209,731

209,731

201,088

201,088

Panel (b):
sum of coefficients

1st

(1)

MP

-0.0201***
(0.0027)

(2)

-

Small banks - by equity ratio quartile

2nd

(3)

-0.0189***
(0.0021)

(4)

-

3rd

(5)

-0.0213***
(0.0029)

(6)

-

4th
(7)

-0.0200***
(0.0035)

(8)

-

Intra*MP

Inter*MP

-0.0014
(0.0036)

-0.0211*
(0.0120)

-0.0017
(0.0029)

-0.0480**
(0.0103)

0.0023
(0.0038)

0.0018
(0.0034)

-0.0020
(0.0045)

-0.0025
(0.0034)

0.0041
(0.0032)

0.0032
(0.0040)

-0.0250** -0.0427***
(0.0101)
(0.0107)

-0.0142 -0.0359***
(0.0099)
(0.0118)

-0.0111 -0.0381**
(0.0104)
(0.0165)

observations

181,399

181,399

200,779

200,779

206,024

206,024

198,825

198,825

Panel (c):
sum of coefficients

1st

(1)

MP

-0.0251***
(0.0038)

Intra*MP

Inter*MP

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

-0.0009
(0.0048)

-0.0182*
(0.0094)

23,341
Yes
Yes
Yes
Yes
-

(2)

-

-0.0007
(0.0040)

-0.0034
(0.0129)

23,341
Yes
-
Yes
-
Yes

Large banks - by equity ratio quartile

2nd

(3)

-0.0251**
(0.0095)

-0.0056
(0.0051)

0.0120
(0.0243)

7,321
Yes
Yes
Yes
Yes
-

(4)

-

-0.0064
(0.0064)

0.0004
(0.0307)

7,321
Yes
-
Yes
-
Yes

3rd

(5)

-0.0149
(0.0135)

(6)

-

-0.0285**
(0.0116)

-0.0278***
(0.0078)

-0.0021
(0.0250)

-0.0067
(0.0398)

3,707
Yes
Yes
Yes
Yes
-

3,707
Yes
-
Yes
-
Yes

4th
(7)

-0.0242
(0.0376)

-0.0111
(0.0098)

0.0603
(0.0511)

2,263
Yes
Yes
Yes
Yes
-

(8)

-

-0.0141
(0.0136)

-0.0166
(0.0731)

2,263
Yes
-
Yes
-
Yes

Panel (a) reports results from separately estimating equation 1.2 for all banks that fall into the 1st,
2nd, 3rd, and 4th quartiles of equity ratio within a given quarter. Panel (b) reports results for small
banks only and panel (c) reports results for large banks only. Equity ratio is measured as total
equity divided by total assets. Robust standard errors clustered at the state-level are in parentheses.
* indicates statistical significance at the 10% level. ** indicates statistical significance at the 5%
level. *** indicates statistical significance at the 1% level.

39

Table 1.12: Market Structure

Panel (a)
Dependent Variable:

Effect of deregulation on market structure

Lerner Index

County HHI

State HHI

-0.0055
(0.0106)

-0.0193*
(0.0104)

3,825
Yes
-
Yes

-

-

-

-

Intra

Inter

observations
State Fixed Effects
Bank Fixed Effects
Time Fixed Effects
Panel (b)
sum of coefficients

MP

LI*MP

County_HHI*MP

State_HHI*MP

Inter*MP

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies

-0.0020
(0.0021)

0.0043*
(0.0023)

853,404

-
Yes
Yes

0.0181***
(0.0018)

0.0079***
(0.0020)

857,525

Yes
-
Yes

Dependent variable: real loan growth

(1)

(2)

(3)

(4)

(5)

(6)

-0.0291*** -0.0287***
(0.0034)
(0.0034)

-0.0204*** -0.0200***
(0.0020)
(0.0024)

-0.0197*** -0.0193***
(0.0023)
(0.0025)

0.0243***
(0.0084)

0.0253***
(0.0081)

-

-

-0.0019
(0.0031)

-0.0012
(0.0031)

-

-

-

-

-

-0.0206**
(0.0085)

-

-

-

-0.0272
(0.0200)

-0.0245
(0.0203)

-0.0205**
(0.0085)

-

-0.0199**
(0.0085)

819,992

819,992

823,659

823,659

823,659

823,659

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Yes
Yes
Yes
Yes

Panel (a) reports the effect of deregulation on the three bank market structure variables: Lerner index
(proxy for market power), county-level HHI (concentration), and state-level HHI (concentration).
Panel (b) reports results from estimating equation 1.7 with the three bank market structure variables.
Odd columns include the interaction between the bank market structure and monetary policy only.
Even columns include the interaction between interstate deregulation and monetary policy as
well. Robust standard errors clustered at the state-level are in parentheses. * indicates statistical
significance at the 10% level. ** indicates statistical significance at the 5% level. *** indicates
statistical significance at the 1% level.

40

Table 1.13: Impact on Lending Response: by Bank Market Power

Panel (a):
sum of coefficients

1st

(1)

MP

-0.0184***
(0.0025)

(2)

-

All banks - by lerner index quartile

2nd

(3)

-0.0201***
(0.0025)

(4)

-

3rd

(5)

-0.0214***
(0.0027)

(6)

-

4th

(7)

-0.0211***
(0.0032)

(8)

-

Intra*MP

Inter*MP

0.0027
(0.0038)

0.0021
(0.0040)

0.0015
(0.0037)

0.0013
(0.0031)

-0.0235*
(0.0129)

-0.0537***
(0.0130)

-0.0197**
(0.0075)

-0.0355***
(0.0123)

0.0006
(0.0030)

-0.0066
(0.0080)

0.0000
(0.0039)

-0.0271**
(0.0107)

-0.0049
(0.0036)

-0.0160
(0.0122)

-0.0055
(0.0033)

-0.0405***
(0.0143)

observations

202,018

202,018

207,650

207,650

207,048

207,048

206,943

206,943

Panel (b):
sum of coefficients

1st

(1)

MP

-0.0182***
(0.0025)

(2)

-

Small banks - by lerner index quartile

2nd

(3)

-0.0199***
(0.0025)

(4)

-

3rd

(5)

-0.0211***
(0.0028)

(6)

-

4th

(7)

-0.0216***
(0.0033)

(8)

-

Intra*MP

Inter*MP

0.0028
(0.0040)

0.0022
(0.0042)

0.0018
(0.0038)

0.0016
(0.0033)

-0.0246*
(0.0129)

-0.0567***
(0.0132)

-0.0200***
(0.0075)

-0.0363***
(0.0124)

0.0003
(0.0032)

-0.0065
(0.0082)

-0.0004
(0.0042)

-0.0286***
(0.0108)

-0.0027
(0.0041)

-0.0180
(0.0125)

-0.0034
(0.0031)

-0.0434***
(0.0150)

observations

199,197

199,197

202,937

202,937

199,059

199,059

185,843

185,843

Panel (c):
sum of coefficients

1st

(1)

MP

Intra*MP

Inter*MP

-0.0300**
(0.0142)

0.0094
(0.0088)

-0.0238
(0.0667)

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

2,821
Yes
Yes
Yes
Yes
-

(2)

-

0.0084
(0.0090)

0.0538
(0.0931)

2,821
Yes
-
Yes
-
Yes

Large banks - by lerner index quartile

2nd

(3)

-0.0304***
(0.0093)

-0.0020
(0.0066)

-0.0206
(0.0221)

4,713
Yes
Yes
Yes
Yes
-

(4)

-

-0.0020
(0.0050)

0.0063
(0.0285)

4,713
Yes
-
Yes
-
Yes

3rd

(5)

-0.0306***
(0.0059)

0.0100*
(0.0053)

-0.0056
(0.0129)

(6)

-

4th

(7)

-0.0177***
(0.0064)

(8)

-

0.0101**
(0.0042)

-0.0043
(0.0152)

-0.0172***
(0.0052)

-0.0172**
(0.0071)

-0.0076
(0.0118)

-0.0148
(0.0153)

7,989
Yes
Yes
Yes
Yes
-

7,989
Yes
-
Yes
-
Yes

21,109
Yes
Yes
Yes
Yes
-

21,109
Yes
-
Yes
-
Yes

This table reports results for estimating equation 1.2 across Lerner index quartile, where the 1st
quartile has the lowest market power and the 4th has the highest market power. Panel (a) reports
results for all banks, panel (b) reports results for small banks (below the 95th percentile in assets),
and panel (c) reports results for large banks (above the 95th percentile in assets).

41

Table 1.14: Impact on Lending Response: by Market Concentration

Panel (a):
sum of coefficients

1st

(1)

MP

-0.0202***
(0.0030)

(2)

-

Intra*MP

Inter*MP

0.0155
(0.0091)

-0.0154
(0.0145)

0.0129**
(0.0064)

-0.0264*
(0.0148)

observations

204,761

204,761

All banks - by HHI quartile

2nd

(3)

-0.0178***
(0.0033)

(4)

-

3rd

(5)

-0.0194***
(0.0022)

(6)

-

4th

(7)

-0.0232***
(0.0026)

(8)

-

-0.0037
(0.0032)

-0.0038
(0.0041)

-0.0227
(0.0042)

-0.0034
(0.0035)

0.0018
(0.0026)

0.0013
(0.0031)

-0.0205*** -0.0411**
(0.0068)
(0.0182)

-0.0207**
(0.0101)

-0.0448***
(0.0133)

-0.0210*
(0.0107)

-0.0406***
(0.0130)

206,920
Small banks - by HHI quartile

206,920

206,337

206,337

205,641

205,641

Panel (b):
sum of coefficients

1st

(1)

MP

-0.0196***
(0.0030)

(2)

-

2nd

(3)

-0.0178***
(0.0035)

(4)

-

3rd

(5)

-0.0191***
(0.0023)

(6)

-

4th

(7)

-0.0238***
(0.0026)

(8)

-

Intra*MP

Inter*MP

0.0111
(0.0089)

-0.0162
(0.0147)

0.0122*
(0.0071)

-0.0307*
(0.0156)

observations

196,482

196,482

-0.0025
(0.0037)

-0.0026
(0.0040)

-0.0021
(0.0046)

-0.0030
(0.0036)

0.0021
(0.0028)

0.0015
(0.0031)

-0.0176*** -0.0373**
(0.0065)
(0.0184)

-0.0207**
(0.0103)

-0.0457***
(0.0132)

-0.0224**
(0.0108)

-0.0430***
(0.0132)

197,437
Large banks - by HHI quartile

197,437

196,086

196,086

197,022

197,022

Panel (c):
sum of coefficients

1st

(1)

MP

-0.0228***
(0.0080)

(2)

-

2nd

(3)

-0.0258***
(0.0064)

(4)

-

3rd

(5)

-0.0262***
(0.0052)

(6)

-

Intra*MP

Inter*MP

0.0008
(0.0187)

-0.0099
(0.0378)

0.0011
(0.0142)

-0.0038
(0.0516)

0.0018
(0.0071)

-0.0312
(0.0200)

0.0019
(0.0071)

-0.0233
(0.0268)

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

8,279
Yes
Yes
Yes
Yes
-

8,279
Yes
-
Yes
-
Yes

9,483
Yes
Yes
Yes
Yes
-

9,483
Yes
-
Yes
-
Yes

-0.0068
(0.0045)

-0.0226
(0.0142)

10,251
Yes
Yes
Yes
Yes
-

-0.0067
(0.0044)

-0.0157
(0.0145)

10,251
Yes
-
Yes
-
Yes

(7)

-0.0099
(0.0094)

-0.0030
(0.0091)

-0.0126
(0.0158)

8,619
Yes
Yes
Yes
Yes
-

4th

(8)

-

-0.0036
(0.0046)

-0.0165
(0.0101)

8,619
Yes
-
Yes
-
Yes

This table reports results for estimating equation 1.2 across county-level HHI quartile, where the
1st quartile has the lowest market concentration and the 4th has the highest market concentration.
Panel (a) reports results for all banks, panel (b) reports results for small banks (below the 95th
percentile in assets), and panel (c) reports results for large banks (above the 95th percentile in
assets).

42

Table 1.15: Loan Portfolio Composition

Effect of deregulation on category share of total loans

C&I Share

-0.0043***
(0.0008)

-0.0064***

(00013)

857,525

Yes
Yes
Yes
Yes

RE Share

-0.0060***
(0.0020)

-0.0174***
(0.0023)

857,525

Yes
Yes
Yes
Yes

Con Share

0.0017
(0.0015)

0.0037*
(0.0022)

857,525

Yes
Yes
Yes
Yes

By loan category

Real Estate Loans
(4)
(3)

Consumer Loans
(6)
(5)

C&I Loans

(2)

(1)

Panel (a)

Intra

Inter

observations
BANK
STATE
Bank Fixed Effects
Time Fixed Effects

Panel (b)

sum of coefficients

MP

-0.0225***
(0.0043)

-

-0.0158***
(0.0031)

-

-0.0379***
(0.0029)

-

Intra*MP

Inter*MP

-0.0141**
(0.0070)

-0.0152***
(0.0055)

-0.0230
(0.0189)

-0.0434**
(0.0194)

0.0024
(0.0038)

-0.0104
(0.0088)

0.0012
(0.0030)

-0.0401***
(0.0102)

0.0020
(0.0037)

0.0046
(0.0104)

0.0007
(0.0045)

-0.0365***
(0.0127)

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

737,753

737,753

795,076

795,076

778,630

778,630

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

Yes
Yes
Yes
Yes
-

Yes
-
Yes
-
Yes

Panel (a) reports the effect of deregulation on the share of total loans for each of the three major
loan categories: commercial and industrial loans, real estate loans, and consumer loans. Panel (b)
reports results from estimating equation 1.2 for the three loan categories. Robust standard errors
clustered at the state-level are in parentheses. * indicates statistical significance at the 10% level.
** indicates statistical significance at the 5% level. *** indicates statistical significance at the 1%
level.

43

Table 1.16: Impact on Lending Response: by BHC Affiliation

All banks - by BHC affiliation

Panel (a)
sum of coefficients

Stand Alone
(1)

(2)

Affiliated

(3)

MP

-0.0236***
(0.0020)

-

-0.0151***
(0.0026)

(4)

-

Intra*MP

Inter*MP

observations

0.0004
(0.0026)

-0.0057
(0.0098)

0.0003
(0.0026)

-0.0018
(0.0057)

-0.0032
(0.0047)

-0.0252**
(0.0125)

-0.0269*** -0.0516***
(0.0098)
(0.0125)

376,569
Small banks - by BHC affiliation

376,569

447,090

447,090

Panel (b):
sum of coefficients

Stand Alone
(1)

MP

-0.0236***
(0.0020)

(2)

-

Affiliated
(3)

-0.0149***
(0.0028)

(4)

-

Intra*MP

Inter*MP

observations

0.0004
(0.0026)

-0.0053
(0.0100)

0.0003
(0.0027)

-0.00003
(0.0072)

-0.0022
(0.0054)

-0.0253**
(0.0126)

-0.0283*** -0.0548***
(0.0125)

(0.0099)

370,452
Large banks - by BHC affiliation

370,452

416,575

416,575

Panel (c):
sum of coefficients

Stand Alone
(1)

Affiliated
(3)

MP

Intra*MP

Inter*MP

observations
STATE
NATIONAL
Bank Fixed Effects
Year Dummies
Time Fixed Effects

-0.0172
(0.0116)

-0.0018
(0.0067)

-0.0394
(0.0266)

6,117
Yes
Yes
Yes
Yes
-

(2)

-

-0.0023
(0.0037)

-0.0389*
(0.0223)

6,117
Yes
-
Yes
-
Yes

-0.0207***
(0.0040)

-0.0062*
(0.0033)

-0.0055
(0.0114)

30,515
Yes
Yes
Yes
Yes
-

(4)

-

-0.0063
(0.0042)

-0.0049
(0.0144)

30,515
Yes
-
Yes
-
Yes

Panel (a) reports results from separately estimating equation 1.2 for stand alone banks and banks
affiliated with a BHC. Panel (b) reports results for small banks only and panel (c) reports results for
large banks. A bank is affiliated with a BHC if they have a direct or regulatory holder identification
number in a given quarter. * indicates statistical significance at the 10% level. ** indicates
statistical significance at the 5% level. *** indicates statistical significance at the 1% level.

44

Table 1.17: Impact on Lending Response: by Size, Liquidity, and BHC Affiliation

Dependent Variable: Real Loan Growth

Panel (a):
Quartile:

Small, BHC Affiliated - by liquidity ratio quartile
4th
1st

2nd

3rd

Inter*MP

-0.0780***
(0.0241)

-0.0683***
(0.0124)

-0.0318**
(0.0141)

-0.0235
(0.0169)

observations
Panel (b):
Quartile:

110,076

106,942

103,685

95,872

Small, Stand Alone - by liquidity ratio quartile
1st

2nd

3rd

4th

Inter*MP

-0.0395***
(0.0139)

-0.0143
(0.0119)

-0.0132
(0.0132)

-0.0265
(0.0181)

observations
Panel (c):
Quartile:

Inter*MP

observations
Panel (d):
Quartile:

94,184

94,483
Large, BHC Affiliated - by liquidity ratio quartile
1st
4th

91,622

2nd

3rd

90,163

-0.0015
(0.0586)

3,045

0.0260
(0.0223)

5,057

0.0178
(0.0131)

-0.0263
(0.0219)

8,615

13,798

Large, Stand Alone - by liquidity ratio quartile
1st

2nd

3rd

4th

Inter*MP

-0.0820
(0.2522)

-0.1319*
(0.0774)

0.0562
(0.0506)

-0.0221
(0.0561)

observations
Bank Fixed Effects
Time Fixed Effects

669
Yes
Yes

1,323
Yes
Yes

2,017
Yes
Yes

2,108
Yes
Yes

This table reports results from estimating equation 1.2 by bank liquidity ratio quartile, with the 1st
quartile being the least liquid and the 4th quartile being the most liquid. Panel (a) reports results
for small banks affiliated with a bank holding company. Panel (b) reports results for small, stand
alone banks. Panel (c) reports results for large banks affiliated with a bank holding company. Panel
(d) reports results for large, stand alone banks.

45

Table 1.18: Balance Sheet Adjustment: by Size, Liquidity, and BHC Affiliation

Dependent Variable: Asset Growth

Panel (a):
Quartile:

Small, BHC Affiliated - by liquidity ratio quartile
4th
1st

2nd

3rd

Inter*MP

-0.0165*
(0.0093)

-0.0168*
(0.0093)

-0.0253**
(0.0100)

0.0197*
(0.0108)

Panel (b):
Quartile:

Small, Stand Alone - by liquidity ratio quartile
4th
1st

2nd

3rd

Inter*MP

-0.0129
(0.0111)

-0.0110
(0.0121)

-0.0012
(0.0124)

-0.0127
(0.0154)

Dependent Variable: Securities Growth

Panel (c):
Quartile:

Small, BHC Affiliated - by liquidity ratio quartile
1st
4th

2nd

3rd

Inter*MP

0.0617*
(0.0371)

0.0698**
(0.0283)

-0.0626**
(0.0298)

0.0773*
(0.0456)

Panel (d):
Quartile:

Small, stand alone - by liquidity ratio quartile
1st
4th

2nd

3rd

Inter*MP

0.0161
(0.0371)

-0.0368
(0.0403)

0.0037
(0.0397)

0.0456
(0.0448)

Bank Fixed Effects
Time Fixed Effects

Yes
Yes

Yes
Yes

Yes
Yes

Yes
Yes

Panels (a) and (b) reports results from estimating equation 1.2 with asset growth as the dependent
variable. Panel (a) reports results for small banks affiliated with a BHC. Panel (b) reports results
for small, stand alone banks. Panels (c) and (d) reports results from estimating equation 1.2 with
securities growth as the dependent variable. Panel (c) reports results for small banks affiliated with
a BHC. Panel (d) reports results for small, stand alone banks. The sample is split by liquidity
quartile, with the 1st quartile including the least liquid banks and the 4th quartile including the
most liquid.

46

Table 1.19: Share of National Lending

Share of Total Loans (National)

Small Affiliated
Small Stand Alone
Large Affiliated
Large Stand Alone

Whole Sample

0.16
0.12
0.61
0.11

1976 - 1985

0.13
0.16
0.60
0.11

1986 - 1994

0.18
0.08
0.62
0.12

This table reports the share of total loans at the national level for four categories of banks: small
banks (below the 95th percentile in assets) affiliated with a bank holding company (BHC), small
stand alone banks, large banks (above the 95th percentile in assets) affiliated with a BHC, and
large stand alone banks. Column 1 presents average share for the entire sample (1976-1994),
column 2 presents average share for the early part of the sample (when the majority of states had
not deregulated), and column 3 presents average share for the later part of the sample (when the
majority of states had deregulated).

47

Table 1.20: Impact on Loan Growth Response: State-Level Results

sum of coefficients

All Banks

Small Affiliated

Small Stand Alone

State-Level Results

Intra*MP

Inter*MP

observations

Intra*MP

Inter*MP

observations
State Fixed Effects
Time Fixed Effects

0.0047
(0.0090)

-0.0189
(0.0399)

3,621

-0.0162
(0.0259)

-0.0801*
(0.0424)

3,539

Large Affiliate

Large Stand Alone

0.0386
(0.0300)

-0.0174
(0.0536)

3,367
Yes
Yes

-0.0613
(0.0617)

0.1382
(0.1206)

1,639
Yes
Yes

0.0241
(0.0266)

0.1372
(0.1092)

3,621

Yes
Yes

This table reports results from estimating equation 1.2 with data aggregated at the state-level. The
dependent variables for each respective column are real loan growth for all loans within a state, real
loan growth for all small BHC affiliated banks within a state, real loan growth for all small stand
alone banks within a state, real loan growth for all large BHC affiliated banks within a state, and
real loan growth for all large stand alone banks within a state.

48

Figure 1.1: Monetary Policy Shock Series

49

Year19761978198019821984198619881990199219941996Percentage Points-8-6-4-20246810Monetary Policy IndicatorsRRFFRFigure 1.2: Real Loan Growth

50

Year19761978198019821984198619881990199219941996Percentage Points-4-3-2-10123456Average Real Loan GrowthFigure 1.3: Bank Credit to Total Private U.S. Credit

51

Year1976197819801982198419861988199019921994Share0.30.320.340.360.380.40.420.44Commercial Bank Loan Share of Total Private CreditFigure 1.4: Average Loan Rate

52

Year19761978198019821984198619881990199219941996Percentage Points0510152025Average Loan RateFigure 1.5: Bank Type Share of National Lending

53

197619781980198219841986198819901992199400.10.20.30.40.50.60.7National Loan Share by Bank TypeSmall Stand AloneLarge Stand AloneSmall BHCLarge BHCFEDERAL RESERVE PRIVATE INFORMATION AND THE STOCK MARKET

CHAPTER 2

Joint work with Aeimit Lakdawala

2.1 Introduction

Uncovering the nature of the monetary policy transmission mechanism continues to be an
important issue in macroeconomics. The large body of literature on this topic has identified a variety
of channels through which monetary policy can affect the economy. However recent research has
emphasized a new so-called “Fed Information" channel whereby a signal from the central bank
that reveals information about economic fundamentals can affect agents’ expectations and thus the
economy (see for example Campbell, Fisher, Justiniano, and Melosi (2016) and Nakamura and
Steinsson (2017)). In this paper we aim to shed light on this Fed information effect through the
lens of the stock market. Specifically, we study how the stock market responds to monetary policy
by explicitly separating exogenous shocks from shocks that reveal information about economic
activity (labeled “Delphic" shocks following Campbell, Evans, Fisher, and Justiniano (2012)). We
focus on the stock market reaction as it is an important component of the overall monetary policy
transmission mechanism which can drive economic activity by affecting wealth, cost of capital and
overall expectations.

We build on the framework of Bernanke and Kuttner (2005) (BK henceforth) and use an
identification strategy based on high-frequency futures market data. Since stock prices should
not react to policy changes that are already anticipated, changes in futures prices that occur in a
narrow window around the Federal Open Market Committee (FOMC) announcements are used to
construct a measure of monetary policy surprise. Given the growing importance of Federal Reserve
communication,1 we extend the federal funds target rate based monetary policy surprise used by
1For early work on the importance of Federal Reserve communication see Gürkaynak, Sack,

54

BK to also include any communication about unexpected future changes in monetary policy. This
is done using an extended set of futures data with a variety of maturities, starting from the current
month up to 4 quarters ahead.

Our estimation methodology proceeds in two steps. First, we decompose the futures based
In
monetary policy surprise measure into an exogenous component and a Delphic component.
carrying out this decomposition we take the view that Federal Reserve signals about economic
activity should surprise the futures market only if they reveal any private information that the Federal
Reserve possesses. This private information could arise either due to asymmetric information
underlying the forecasts or due to a difference in forecasting models. It is important to point out that
our framework does not require that the Federal Reserve actually has superior information relative
to the market ( `a la Romer and Romer (2000)). In other words, even in a hypothetical case where
the market knows its forecast is more accurate, the Federal Reserve’s forecast may be useful to the
market since policy actions will be based on that forecast.

To capture this private information we construct a measure that combines market survey data
with the Federal Reserve’s internal forecasts. Specifically, our measure is defined as the difference
between the Greenbook forecasts produced by the Federal Reserve Board’s staff and the consensus
forecast from the market based Blue Chip survey. The first step involves running a regression of
monetary policy surprise on our measure of private information. The regression results suggest
that monetary policy surprises are “predictable" using the private information variable. Moreover,
the estimates imply that when the Greenbook forecast is more optimistic relative to the market’s
forecast, it is related to a positive monetary policy surprise (i.e. a contractionary surprise). Note
that this regression can only be run ex-post and not in real-time as the Greenbook forecasts are
publicly released with a five year lag. But the statistically significant and systematic relationship
between the monetary policy surprise and Federal Reserve private information suggests that a
portion of the futures market reaction is attributable to differences in forecasts.2
and Swanson (2005). For a more recent study see Feroli, Greenlaw, Hooper, Mishkin, and Sufi
(2017).

2We would like to point out that our measure of private information may not perfectly capture

55

The second step in our estimation involves studying the stock market response to the fitted
value (i.e. Delphic component) and the residual (clean measure of an exogenous monetary policy
shock) of the first step regression. We layout a simple conceptual framework to understand how
exogenous and Delphic shocks can affect stock prices differently. Under some simple conditions,
the stock response to an exogenous shock is expected to be negative, while that of the Delphic shock
can be either positive or negative. Our baseline results using data from 1991 to 2011 find a stock
response to the exogenous monetary policy shock that is similar to BK.3 A hypothetical surprise
increase of 100 basis points in the expected path of the fed funds rate over the next 4 quarters
results in about a 5.7% fall in the S&P 500 index. On the other hand, a contractionary Delphic
shock of the same size reduces stock prices by about 2.1%; a statistically significant difference
relative to the exogenous monetary policy shock response. Thus, on average stock prices fall in
response to surprise contractionary shocks, whether they are exogenous or Delphic in nature. But
we also find some evidence for an asymmetry in the stock response on certain FOMC meetings,
especially concerning Delphic shocks. These episodes occur when FOMC policy actions were
enacted at unscheduled dates (also called inter-meeting moves) or when there is a reversal in the
direction of the change in the fed funds rate target (also called turning points). On these particular
FOMC meetings, the stock market falls more in response to a contractionary exogenous monetary
policy shock but actually rises in response to a contractionary Delphic shock. Previous studies
have found differential effects of monetary policy shocks on unscheduled FOMC meetings (BK and
Faust, Swanson, and Wright (2004a) among others). Our results suggest that this is partly due to
the Delphic component of the monetary policy surprise at these meetings that has been previously
unexplored in the literature.

To complement our high-frequency analysis and to understand the economic reasons behind the
observed stock price response, we perform a decomposition of stock prices using the framework
the true underlying information differences. We would like to observe both the FOMC members’
and the market’s macro forecasts closer to the FOMC announcement (ideally just a few minutes
before the announcement). Unfortunately, data at this frequency is not available and we believe our
measure using the Greenbook and Blue Chip forecasts is the best proxy based on existing data.

3The end date of the sample is restricted by the most recently available Greenbook data.

56

of Campbell and Ammer (1993). This methodology uses a monthly vector autoregression to break
down current excess stock returns into revisions of the expectation of discounted future dividends,
the real interest rate, and future excess returns. We find some suggestive evidence that on average the
response of excess returns to exogenous shocks is mostly due to changes in expected future excess
returns and dividends, while the excess return response to a Delphic shock is primarily attributed
to changes in expected dividends. These vector autoregression results confirm the asymmetric
effects of monetary policy actions (especially the Delphic shocks) on unscheduled and turning
point FOMC meetings. The stock response to Delphic shocks on these meetings appears to be
driven mostly by movements in the expected future excess returns.

This paper lies at the intersection of two distinct strands of the literature. First, there is a long
line of work that builds on the high frequency approach of BK to study the effect of monetary
policy on stock prices. Gürkaynak, Sack, and Swanson (2005) and more recently Kurov (2012) and
Eijffinger, Mahieu, and Raes (2017) expand on this work by separately estimating the stock response
to surprises to the federal funds rate and surprises in forward guidance. There has also been work
exploring the cross-sectional firm level stock price reactions to monetary policy, see Gorodnichenko
and Weber (2016), Ehrmann and Fratzscher (2004), Ippolito, Ozdagli, and Perez-Orive (2013),
Maio (2013), Jansen and Tsai (2010) and Laeven and Tong (2012) among others. While our
analysis focuses exclusively on a narrow window around FOMC meetings, there is intriguing new
evidence that discusses other occasions on which the Federal Reserve communicates to the public
(for example in speeches made by FOMC members). These are explored in more detail by Cieslak,
Morse, and Vissing-Jorgensen (2016), Lucca and Moench (2015) and Neuhierl and Weber (2017).
However, all these papers use the composite monetary policy surprise measure, while the focus of
our paper is to separate the effect of Delphic monetary shocks from the exogenous monetary policy
shocks. Second, this paper is also related to the growing literature on how central bank signals
about fundamentals can affect economic activity. Campbell, Fisher, Justiniano, and Melosi (2016)
and Nakamura and Steinsson (2017) empirically highlight the role of Delphic signals and their
effect on survey expectations. Melosi (2016) and Tang (2015) provide evidence of this channel

57

using a dynamic stochastic general equilibrium model while Lakdawala (2017) uses a structural
vector autoregression framework. The stock market response results in this paper are consistent
with this literature. In light of the growing evidence of the signaling channel of monetary policy,
we advocate accounting for the asymmetric effects on these meetings to get the full picture of the
monetary transmission mechanism.

2.2 Stock Prices and Monetary Policy

To identify the effect of monetary policy on stock prices, one cannot directly regress stock
prices on the central bank’s policy instrument (for example the short-term interest rate). The
endogenous reaction of both stock prices and the central bank’s policy instrument to common
economic conditions leads to the classic simultaneous equation bias. Thus the literature has tried
to isolate exogenous variation in the policy instrument to overcome this problem. Following
the work of BK, an important strategy involves high-frequency identification using federal funds
futures contracts. In this section we first outline a simple framework to understand futures based
identification, with a special emphasis on why central bank private information can matter. This
treatment is closely related to the framework laid out in Miranda-Agrippino (2016). Next we extend
the framework and discuss how stock prices may respond differently to an interest rate change by
the central bank depending on if the change reflects an exogenous monetary policy shock or if it
reflects a signal about the central bank’s private information.

2.2.1 Monetary Policy Surprise from Futures Data
Let p(h)
this futures contract is the federal funds rate.4 Thus we can write

t

be the price of a futures contract at time t that matures in t + h. The underlying asset for

(h)
p(h)
t = it+h|t + ζ
t

(2.1)

4Note that technically the fed funds futures contract trades as 100 - the average effective fed

funds rate, but we are omitting the "100 -" component for simplicity.

58

(h)
where it+h|t = Etit+h is the expected fed funds rate at t + h and ζ
is the risk-premium. There is
t
an ongoing debate in the literature about the relevance of risk-premia in fed funds futures markets,
but they are not crucial to our analysis and we will set them to zero in the illustrative model.5

The next step is to consider a general monetary policy rule where the central bank changes
the short-term interest rate it in response to current, lagged and forecasts of certain indicators of
economic activity.

+ et

(cid:98)ΩCB

(2.2)
where et represents a monetary policy shock and gCB(.) is the central bank’s reaction function.
contains the central bank information set available at time t, including any current information
t|t
that is used to form forecasts. The hat denotes the fact that the central bank estimates the values of
the relevant variables based on their information set.6

it = gCB(cid:16)(cid:98)ΩCB

t|t

(cid:17)

An important convention in the monetary policy literature is that et is assumed to be an
exogenous shock, i.e. it is unrelated to economic activity. Thus if we can estimate et, then we can
regress stock prices on et to identify the effects of monetary policy. One strategy for identification
is to study changes in fed funds futures data around FOMC announcements, following BK.

Consider the futures contract maturing at the end of the current month (i.e. h = 0). Specifically,

consider the futures prices of this contract measured just before the FOMC announcement

(cid:17)

(cid:16)(cid:98)ΩM

t|t

p(0)
t−ε

= it|t−ε

= g

(2.3)

market’s information set,(cid:98)ΩM

The M superscript denotes the fact that the futures price will reflect expectations based on the
. We are making the assumption that the market has full knowledge
of the central bank’s reaction function, i.e. gCB(.) = gM(.) = g(.). Below, we provide an alternate
derivation of our estimating equation where we relax this assumption.

t|t

5Piazzesi and Swanson (2008) find that fed funds futures risk-premia are slow-moving and do
not change much around FOMC announcements. On the other hand, Miranda-Agrippino (2016)
finds a bigger role for risk-premia.

denotes the period k estimates of the fundamentals in the monetary policy

6In general Ω

j
k|t

reaction function based on the information available to j in period t.

59

The key assumption in the futures based identification is that no other macro news announce-

ments are released in the window between t − ε and t. Thus we have that(cid:98)Ωt|t−ε

=(cid:98)Ωt|t. Now

consider the futures price after the FOMC announcement.

(cid:17)

(cid:16)(cid:98)ΩCB

t|t

p(0)
t = it|t = g

+ et

(2.4)

Note that the information set that is relevant to the short rate set by the central bank is its own
information set. The monetary policy surprise is measured as the change in the futures contract

mpst = p(0)
= g

= g

(cid:17)

(cid:16)(cid:98)ΩM
(cid:16)(cid:98)ΩCB
(cid:17) − g
t − p(0)
t−ε
(cid:16)(cid:98)ΩCB
(cid:17)
t|t −(cid:98)ΩM

t|t

t|t

t|t
+ et

+ et

(2.5)

(cid:17). The price
= gM(cid:16)(cid:98)ΩM
(cid:17)

t = it|t = gCB(cid:16)(cid:98)ΩCB

where the last equality holds if we assume a linear reaction function g(.) for the central bank. There is
an alternative way to derive equation 2.5 without resorting to the assumption that the market has full
knowledge of the central bank’s reaction function. In this case we will assume gCB(.) to represent
the central bank’s actual monetary policy stance given its estimates of the relevant fundamentals,
rather than just the reaction function component of its rule, i.e. p(0)
of the futures contract just before the FOMC announcement is given by p(0)
t−ε
t|t
where gM(.) is not assumed to be the same as gCB(.). Then if gCB(.) and gM(.) are linear we can
write the monetary policy surprise as
t − p(0)
t−ε
t|t

(cid:17)
(cid:17) − gM(cid:16)(cid:98)ΩM
= gCB(cid:16)(cid:98)ΩCB
(cid:17)
(cid:17) − gCB(cid:16)(cid:98)ΩM
(cid:17) − gM(cid:16)(cid:98)ΩM
= gCB(cid:16)(cid:98)ΩCB
(cid:17)
= gCB(cid:16)(cid:98)ΩCB
+ gCB(cid:16)(cid:98)ΩM
(cid:17) − gM(cid:16)(cid:98)ΩM
t|t −(cid:98)ΩM
(cid:17)
= gCB(cid:16)(cid:98)ΩCB
t|t −(cid:98)ΩM
In this case the exogenous monetary policy shock et ≡ gCB(cid:16)(cid:98)ΩM

+ gCB(cid:16)(cid:98)ΩM
(cid:17)
(cid:17) − gM(cid:16)(cid:98)ΩM

(2.6)

(cid:17) has a specific

mpst = p(0)

t|t
= it|t−ε

t|t
t|t

t|t
t|t

t|t
t|t

(cid:17)

t|t

t|t

t|t

+ et

t|t

t|t

interpretation and represents the central bank and the market translating the same fundamentals
into different monetary policy stances. On the other hand, the exogenous monetary policy shock

60

from 2.5 represented a broader and more conventional measure of an exogenous monetary policy
shock.

Regardless of the approach taken to derive equation 2.5 (or 2.6), it is clear that in the special
case that the information set of the central bank and the market coincide, the monetary policy
surprise recovers the exogenous monetary policy shock. However the assumption of no asymmetric
information may not be tenable. There is a growing body of literature suggesting a role for central
bank signals about macro fundamentals. Nakamura and Steinsson (2017) find a “Fed information
effect" where Fed communication affects agents’ expectation of future economic activity. Melosi
(2016) sets up a DSGE model with an explicit signalling channel of monetary policy and finds that
it has empirically relevant effects. Finally, Tang (2015) also finds that the empirical patterns in the
U.S. inflation data are consistent with the existence of a signalling channel. While this a nascent
literature, it does seem to suggest that the “signalling/information" channel is important. In this
paper we add to this literature by studying the response of the stock market and testing whether it
responds differently to Delphic shocks when compared to traditional exogenous monetary policy
shocks.

While the derivation presented above used the futures contract expiring in the current month, we
can show more generally that the analysis used to derive equation 2.5 (or 2.6) also applies to futures
contracts that expire not in the current month, but in the future. These surprises likewise capture an
exogenous component, which is a signal about shocks to the interest rate that are expected to occur
in the future. But the surprises also capture a signal about future shocks to the interest rate that
are related to central bank private information about macroeconomic fundamentals (i.e the Delphic
shocks).

In the first step of the estimation procedure we separate the monetary policy surprises into i)
exogenous component and ii) private information component. Equation 2.5 suggests that a simple
linear regression will suffice as long as we can construct a variable that measures the difference
in the information set of the central bank relative to the market. Essentially we need a private
. In section 2.3.2 below we discuss in detail how we

information variable that captures(cid:98)ΩCB

t|t −(cid:98)ΩM

t|t

61

create this variable using forecast data. With this variable in hand, we run the following regression.

Using this equation we construct the residual(cid:98)et and the fitted value(cid:98)γ

mpst = c + γ

+ et

(cid:16)(cid:98)ΩCB
t|t −(cid:98)ΩM

t|t

(cid:17). In the next

(2.7)

(cid:16)(cid:98)ΩCB
t|t −(cid:98)ΩM

t|t

(cid:17)

step of the estimation procedure we regress the change in the stock price on the residual and fitted
value.

∆St = α + β1(cid:98)et + β2(cid:98)γ

(cid:16)(cid:98)ΩCB
t|t −(cid:98)ΩM

t|t

(cid:17)

+ ut

(2.8)

where St is the stock price and ∆ represents the change in a narrow window around the FOMC
announcement. What should we expect for the sign of the two coefficients β1 and β2? Next we
layout a simple “model-free" theoretical framework that can help us understand the related issues.

2.2.2 Stock Price Response to Exogenous and Delphic Shocks

Here we provide the key intuition of how the two different shocks can affect stock prices through
their effects on discount rates and cash-flow news. In the online appendix we provide a conceptual
framework where this intuition is fleshed out in more detail.

Consider a surprise increase in the interest rate by the central bank that is solely due to an
exogenous monetary policy shock. In conventional models where monetary policy has real effects
this translates to bad news about future cash flows and higher discount rates. Thus both discount
rates and cash flow news work to create a fall in stock prices. This is the traditional channel of how
monetary policy affects the stock market and suggests that β1 from equation 2.8 above should be
negative.

But monetary policy can have an additional effect if the change in the interest rate is related to
revelation of central bank private information. A surprise increase in the interest rate in this case
can have an ambiguous effect on stock prices, because there are distinct and potentially opposing
effects. First, consider the effect of a contractionary shock on expectations of future cash flows.
If the rise in interest rates has a contractionary effect on the economy, it will mean bad news
about future cash flows. However this decision to increase interest rates could be driven by the

62

central bank’s forecast being more optimistic relative to the market. This Delphic signal could lead
the market to revise their expectations of economic activity upwards in response. There is some
recent empirical work suggesting that central bank signals can directly affect private sector beliefs
about future economic activity. Melosi (2016) builds a model with an explicit signaling channel
of monetary policy. The model incorporates a mechanism that could lead agents to expect higher
inflation in response to a signal tied to an increase in the interest rate. In a similar vein, Nakamura
and Steinsson (2017) sketch a model where the central bank can affect the market’s expectations
about the natural rate of interest.
In their model an increase in the interest rate can cause the
market to revise upwards their expectation of the natural rate, leading to a rise in economic activity.
Finally, in a recent paper Campbell, Fisher, Justiniano, and Melosi (2016) use similarly constructed
private information variables and show that the component of the monetary policy surprises that
is related to optimistic Fed private information predicts upward revisions of economy activity by
forecasters. This upward revision of expectations will mean good news about future cash flows.
Thus a contractionary Delphic shock can be expected to raise stock prices through its effect on
future cash flows. However, since it is a contractionary shock we would expect it to raise discount
rates and thus lower stock prices. This latter effect goes in the opposite direction of the effect that
works through cash flows, leaving the overall sign ambiguous.

To summarize, the conceptual framework suggests that we should have a strong prior for β1
to be negative but there is uncertainty about the sign of β2 as it can reasonably be expected to be
either positive or negative.

2.3 Data

We use the S&P 500 index to measure the response of the stock market. The prices are
measured in a 30 minute window around FOMC announcements, starting at 10 minutes before the
announcement and ending 20 minutes after the announcement. For our baseline results, we use
the sample period 1991-2011. There are 188 total FOMC policy decisions over this time frame.
We drop a total of five data points. We exclude 8/17/2007, 11/25/2008, and 4/27/2011 due to data

63

unavailability for those dates. We also drop 9/17/2001 and 3/18/2009 following Campbell, Fisher,
Justiniano, and Melosi (2016). This leaves 183 observations in our sample. In the next subsection
we detail the construction of the monetary policy surprise and conclude this section by discussing
the private information variables constructed from Greenbook and Blue Chip forecasts.

2.3.1 Monetary Policy Surprise

Our measure of the surprise change in monetary policy is constructed from interest rate futures
contracts, as in Kuttner (2001). Federal funds rate and Eurodollar futures contracts capture the mar-
ket’s expectations about future Federal Reserve actions. Changes in these futures contracts around
FOMC announcements therefore serve as a measure of the change in policy that is unanticipated
by the market. Since any expected change in policy will already be priced into financial assets, the
reaction of asset prices to monetary policy should be entirely due to this surprise component.

We want the monetary policy surprise measure to capture surprises to expectations about future
fed funds rate changes, in addition to any surprise to the current month’s fed funds rate target.
Thus to construct our measure of the monetary policy surprise, we follow Gürkaynak, Sack, and
Swanson (2005) and use five futures contracts: the current month’s fed funds futures, the 3-month
ahead fed funds futures, and the 2-quarter, 3-quarter, and 4-quarter ahead Eurodollar futures.7 For
the baseline results, the surprise in each contract is measured as the change in the futures rate in a
30 minute window (10 minutes before to 20 minutes after) around FOMC policy decisions as in
Gürkaynak, Sack, and Swanson (2005). But we also discuss results obtained using a broader daily
window. Taken together, the five contracts contain rich information about the short and medium
term path of expected interest rates.

To summarize this information in a parsimonious way we perform a principal components
analysis. Let X denote a T x 5 matrix of the change in the price of the 5 futures contracts, where
T is the number of FOMC meetings. We can then perform a principal components analysis of the
7 For comparison, Bernanke and Kuttner (2005) use only the current month fed funds futures

contract in their baseline results.

64

futures price changes

X = FΛ +(cid:101)η

where F are factors, Λ are factor loadings, and(cid:101)η is an error term. The first principal component

(2.9)

of F explains more than 80% of the total variation across all the contracts. We therefore use this
first principal component as our baseline measure of monetary policy surprises.8 Figure 2.1 plots
this monetary policy surprise measure using both the 30 minute and daily window. The two series
display a high degree of correlation with some minor discrepancies around the financial crisis in
2008 and in the early 1990s. To facilitate interpretation of our results below, we normalize the
policy surprise such that its effect on the four quarter ahead Eurodollar futures contract is equal to
unity. Thus the coefficient from a regression of stocks on the monetary policy surprise will measure
the effect on the stock market of a 1% surprise rise in the expected path of the fed funds rate over
the next 4 quarters.

2.3.2 Federal Reserve Private Information

Our measure of Federal Reserve private information is constructed using the FOMC Greenbook
forecasts and the private sector Blue Chip forecasts, and is similar to the approach used in Barakchian
and Crowe (2013) and Campbell, Fisher, Justiniano, and Melosi (2016).

The Fed’s Greenbook forecasts represent the information set of the central bank(cid:98)ΩCB
equation 2.7, while the Blue Chip forecasts proxy for the market’s information set(cid:98)ΩM

from
. Greenbook
forecasts are constructed by the Federal Reserve Board’s staff a week prior to every scheduled
FOMC policy meeting and are released to the public following a roughly five year lag. Blue Chip
forecasts are compiled from market professionals on a monthly basis and released on the 10th
of every month. For each FOMC policy decision (t) the corresponding measure of Fed private
information is calculated as the most recent Greenbook forecast minus the last Blue Chip forecast
prior to the policy decision that is of the same forecast horizon as the relevant Greenbook forecast.
8This is essentially identical to the measure used in Nakamura and Steinsson (2017) which they

t|t

t|t

call the “policy news shock"

65

In table 2.1, for each FOMC meeting we list the corresponding Greenbook and Blue Chip forecast
dates.

Each set of forecasts predicts the values of macroeconomic variables on a quarterly basis. For
the 1991-2011 sample we use the following four variables: real GDP, CPI, industrial production,
and the civilian unemployment rate. For each variable, both set of forecasts contain at least five
different forecast horizons:
the current quarter forecast, the quarter ahead forecast, two quarter
ahead forecast, three quarter ahead forecast, and four quarter ahead forecast. Our measure of
private information for variable i at forecast horizon j is:

(cid:98)ΩCB
i,t+j|t −(cid:98)ΩM

i,t+j|t

(2.10)

These variables are plotted in figure 2.2. A few interesting points stand out. These variables are
persistent and for each variable as the forecast horizon increases, the persistence rises. This suggests
that the Federal Reserve’s internal forecasts are not completely inferred by the market based on
FOMC meeting actions and announcements. This is especially true for the longer-horizon forecasts.
For a given variable, in addition to the autocorrelation for each individual forecast horizon, the
private information variables for different horizons are also correlated with one another. Forecast
horizons that are “closer" to each other are more highly correlated. For example, the 4 quarter
ahead forecast is quite highly correlated with the 3 quarter ahead forecast but not with the nowcast.
These patterns guide us in choosing the private information measures that will be used in the
regression analysis below. First, given the high cross-correlation among the private information
variables of different horizons (for a given variable) we use only the nowcast and the 4 quarter
ahead forecast. Next, given the high persistence of the private information variables, we include
the first lag in our regression. Thus our baseline specification will have the contemporaneous
and first lag of the nowcast (0 quarter ahead forecast) and 4 quarter ahead forecast for four macro
variables: GDP, CPI, Industrial Production and Unemployment. Thus we have a total of 16 private
information variables that capture the relevant information. A potential alternative is to follow the
approach of Campbell, Fisher, Justiniano, and Melosi (2016) and construct a short and long factor

66

for each variable using principal component analysis. We found that the short factor and long
factors correlate very highly with the nowcast and the 4 quarter ahead forecasts.

2.4 Results

2.4.1 Stock Prices and Monetary Policy Surprise

We start by exploring the relationship between changes in the S&P 500 index (∆St) and our measure
of monetary policy surprise (mpst) detailed in the previous section. Table 2.2 reports the summary
statistics for these two measures using both a tight window and broad window around FOMC
announcements. The tight window measures the change from 10 minute before to 20 minutes after
the announcement. The broad window is just the daily change. The correlation between the tight
and broad measures of the monetary policy surprise is high (0.81), while the correlation is lower
for stock returns (0.47). For the policy surprise, moving to a broader window increases the standard
deviation slightly, but it does so considerably more for the stock return. Thus stock returns in the
broad window appear to have more noise relative to the tight window. The table also provides
information separated by unscheduled FOMC meetings and meetings that correspond to “turning
points" (which are instances when the federal funds rate target is changed in the direction opposite
to previous changes). The specific dates for the unscheduled and turning point FOMC meetings are
listed in table 2.1. There has been some discussion in the literature that FOMC meetings of these
two types are “unusual" relative to the other meetings. BK document that stock price reactions
are much larger on turning point FOMC meetings. Faust, Swanson, and Wright (2004a) find that
monetary policy surprises on unscheduled FOMC meetings are more likely to reveal information
about the state of the economy, i.e. suggesting a role for Delphic shocks (using our terminology).
Finally, using a regime-switching model Davig and Gerlach (2006) show that for a “high volatility"
regime from 1998 to 2002 the effect of monetary policy surprises on stock prices is mainly driven
by unscheduled meetings. We will discuss the importance of these particular episodes for stock
prices in more detail below and in section 2.4.3. For now, we want to point out that all three papers
use data up to the early 2000s (2002 for BK and 2003 for Faust, Swanson, and Wright (2004a) and

67

Davig and Gerlach (2006)). Extending the data up to 2011, we notice that both monetary policy
surprises and stock returns are substantially more volatile on unscheduled and turning point days,
consistent with the idea that these meetings are somewhat different.

Table 2.3 presents the results from the regression of ∆St on mpst using the 30 minute window
with robust standard errors in parentheses. R2
> 0.3 provides support for the assumption that
monetary policy surprises are major drivers of stock prices in this narrow window. Consistent
with BK, the specification in column (1) reports a significant decline in the S&P 500 following
a positive monetary policy surprise (i.e. an unexpected tightening of monetary policy). A 1%
surprise rise in the expected path of the fed funds rate over the next 4 quarters results in a 5.1%
fall in stock prices.9 This coefficient is precisely estimated with statistical significance at the 1%
level.10 Column 2 presents regression results where the monetary policy surprise is interacted with
a dummy variable that jointly represents FOMC meetings that are unscheduled and those associated
with turning points. Column 3 and 4 presents the interaction results where the dummy variable is
separated into unscheduled meetings and turning point meetings. The stock response to a monetary
policy surprise is slightly lower in columns 2-4. The interaction coefficients are all negative but
none of them are statistically significant. These negative point estimates suggest that if there is any
evidence of asymmetry in the response of stock prices, it points to a larger negative response on
unscheduled and turning point FOMC meetings. Since the standard errors are relatively large, it
is reasonable to conclude that the response of stock prices to monetary policy surprises is stable
across these different types of FOMC meetings.

Table 2.4 shows the regression results using the wider daily window. Column 1 shows that the
response of stock prices is now statistically insignificant and much lower in magnitude relative to
the tight window (-2.4% vs −5.1%). The R2 is also substantially lower at .03. The daily stock
9Notice from table 2.2 that the standard deviation of the policy surprise is 7 basis basis points.
This implies that a one standard deviation increase in the policy surprise leads to a 0.36% fall in
stock prices.

10Our results are more strongly significant compared to studies that only use the current month
federal funds futures contract in calculating their monetary policy surprise (see for example Gorod-
nichenko and Weber (2016)).

68

response in table 2.4 is also lower relative to the findings in BK. There are two main reasons
why our daily results are different from BK’s daily results. First, we use a broader measure of
monetary policy surprise that captures forward guidance shocks, while BK just used federal funds
rate surprises. And second, we extend the sample end date from 2002 to 2011. Similar to table
2.3, columns 2-4 show the regression results with dummy interactions for unscheduled and turning
point FOMC meetings. The coefficients on the interactions are negative and two out of the three are
not significant. Thus the daily data regressions confirm that the stock market response to monetary
policy surprises is stable across the different FOMC meetings and if anything more likely to be
negative in theses episodes.

Taken together, it is an indication that stock returns in the broad window have a lot more noise
relative to the tight window. The underlying identifying assumption in this paper is that the relevant
window around FOMC announcements does not contain any other important macroeconomic news
event.
In light of the above results, this identifying assumption is more credible with the tight
window and motivates us to use the tight window for our benchmark results below in section
2.4.3. This is also consistent with the recommendation of Gürkaynak, Sack, and Swanson (2005)
among others. To conclude this section, figure 2.3 shows a scatter plot of the stock return and the
monetary policy surprise in the tight 30 minute window (which is our preferred measure that is
used in the results below). There is a clear negative relationship. The black triangles mark the
Unscheduled FOMC meetings while the red squares represent turning points, highlighting that the
bigger monetary policy surprises occur at these two types of meetings.

2.4.2 Monetary Policy Surprise and Private Information

In section 2.3.2 we discussed the properties of the private information variables constructed from
forecast data. An important implication was that the Federal Reserve does not seem to completely
reveal all of its private information through the FOMC announcement. Thus we would like to
use only the component of private information that is inferred by the market from the FOMC
announcements. As discussed above, we proceed by first regressing the monetary policy surprise

69

measure on the private information variables. The estimating equation is reproduced below

(cid:16)(cid:98)ΩCB
i,t+j|t −(cid:98)ΩM

i,t+j|t

(cid:17)

mpst = c + γi, j

+ et

(2.11)

Table 2.5 shows the results from this regression using the nowcast and 4 quarter ahead forecasts
for the GDP, CPI, Unemployment and Industrial Production private information variables. Given
the persistent nature of the private information variables, we also include the first lag. The p-value
jointly tests the null hypothesis that the private information variables have no explanatory power.
This is rejected at the 1% level. The R2 from the regression is 0.16, which is substantial but also
highlights the fact that a major part of the monetary policy surprise is exogenous with respect to
the Fed’s private information.

variables suggest that an optimistic forecast results in a positive value for(cid:101)gt ≡(cid:98)γ((cid:98)ΩCB

In the conceptual framework sketched out in section 2.2.2, we emphasized that the response of
stock prices to private information depends on how forecast differences are related to interest rate
changes. The regression coefficients from table 2.5 can inform us about the sign. Note that a positive
value for the private information variable for GDP, CPI and IP means that the Fed has a relatively
optimistic forecast for the economy. For unemployment a positive sign implies the opposite. The
first step regression is reported in table 2.5, where 0Q refers to the nowcast and 4Q refers to the
four quarter ahead forecast. The sign of all the coefficients on the private information nowcast
t|t), i.e. a
contractionary policy surprise. But not all the signs on the lagged variables have the signs consistent
with this interpretation. For example, the coefficient on the lagged 4 quarter ahead forecast of IP
implies that if the Fed has a more positive outlook for IP, that is related to an expansionary policy
surprise. This is most likely a combination of some noise and the fact that there is a high amount
of correlation in the content of the different private information variables. We have also run the
first step regression with different combinations of private information variables (including using

t|t −(cid:98)ΩM

principal component analysis) and find that most of the coefficients are consistent with(cid:101)gt being

positive.

Figure 2.4 displays the exogenous monetary policy shock (residual) and Delphic shock (fitted
value) over time, with summary statistics reported in table 2.6. The Delphic shock is typically of

70

a smaller magnitude with a standard deviation roughly half that of the exogenous monetary policy
shock. The standard deviation of the Delphic shock is roughly stable even when we narrow down
to unscheduled or turning point FOMC meetings. On the other hand, the standard deviation of
the exogenous monetary policy shock is much larger in these particular episodes. The Delphic
shock displays a few notable episodes, with relatively large contractionary shocks in the late 90s
and expansionary ones in the early 2000s and 2008-2009. The overall pattern of the exogenous
monetary policy shock is similar to the monetary policy surprise, which is unsurprising given that
the exogenous monetary policy shock explains around 80% of the variation of the monetary policy
surprise.

2.4.3 Stock Price Response to Exogenous and Delphic Shocks

Now we are ready to run our second step regression. We regress the change in the S&P 500 index in
the 30 minute window on the exogenous and Delphic shocks obtained from the first step discussed
above. The estimating equation is

∆St = α + β1(cid:98)et + β2(cid:98)γi, j

(cid:16)(cid:98)ΩCB
i,t+j|t −(cid:98)ΩM

i,t+j|t

(cid:17)

+ ut

(2.12)

Since the regressors in this second step are generated in the first step, we have to account for the
added sampling uncertainty. This is done by bootstrapping the standard errors. The key idea is to
conduct the resampling at the beginning and thus to perform both steps of the two-step regression
procedure for every bootstrap sample. We use 10, 000 replications in the bootstrap procedure.

The results are presented in table 2.7 with the bootstrapped standard errors in parentheses.
Column 1 shows that the exogenous shock has a negative and significant effect on stock returns
with a slightly larger magnitude than the monetary policy surprise. Specifically, a 1% surprise rise
in the expected path of the fed funds rate over the next 4 quarters results in a precisely estimated
5.7% fall in stock prices (relative to the 5.1% fall for the monetary policy surprise).11 The effect of
11The standard deviation of the exogenous shock is slightly lower relative to the monetary policy
surprise. Thus the stock response to a one standard deviation exogenous monetary policy shock is
essentially identical to the monetary policy surprise response.

71

the Delphic shock is also negative but much lower at -2.1%. While this coefficient by itself is not
statistically significant, it is significantly different from the coefficient on the exogenous monetary
policy shock (with a p-value for the difference of 0.05). As shown in table 2.6, exogenous monetary
policy shocks are more volatile than Delphic shocks and we reinterpret the coefficients to get a
better gauge of the size of the effects. Specifically, stock prices fall 0.34% and 0.06% in response to
a one standard deviation exogenous monetary policy and Delphic shock respectively. An important
implication is that on average surprise Federal Reserve decisions and announcements that are
related to revelation of their private information have a lower effect (in terms of both economic and
statistical significance) on the stock market as compared to actions that are exogenous shocks.

However, there is important asymmetry in the effect of these shocks. The second column shows
the results where the exogenous monetary policy and Delphic shocks are interacted with a dummy
variable that jointly represents FOMC meetings that are either unscheduled or are associated with
turning points. The overall stock response is lower in magnitude for the exogenous monetary policy
shock and higher in magnitude for the Delphic shock by about a percentage point. The interaction
coefficient on the exogenous component is −3.6 implying that the total response of stock prices to
exogenous monetary policy shocks on these particular FOMC meetings is substantially larger in
magnitude at −8.2. With a higher variance of exogenous monetary policy shocks on these particular
meetings, stock prices fall by 0.9% in response to a one standard deviation exogenous monetary
policy shock. The interaction coefficient on the Delphic component is also large but positive at 14.6
(with a p-value of 0.04). The total response of stock prices to a Delphic shock on these particular
FOMC meetings is 11.7 (with a p-value of 0.08). Even with a lower variance of Delphic shocks
on these FOMC meetings, it implies a 0.35% rise in stock prices in response to a one standard
deviation Delphic shock. Moreover, the difference between the total effect of exogenous monetary
policy and Delphic shocks for these days is large and strongly significant (p-value < 0.01). Thus
there appears to be a clear distinction in how the stock market interprets exogenous vs. Delphic
monetary policy actions on these particular FOMC meetings.

The third and fourth columns show the results where the interaction for unscheduled FOMC

72

meetings and turning point FOMC meetings is done separately. The same pattern is obtained
with the interaction coefficients. Clearly, the standard errors are larger as there are a total of 17
observations for the unscheduled dates and only 8 for the turning point dates. Nevertheless the sign
of the interaction coefficients on these particular dates continue to show a larger negative response
to the exogenous monetary policy shock and a positive response to the Delphic shock.

Next we check the robustness of the results to sample selection. First, we consider the zero lower
bound episode. Since late 2008, the fed funds rate has been stuck around zero and all the variation in
our monetary policy surprise measure is driven by forward guidance surprises rather than any target
rate change surprise. Furthermore, after hitting the zero lower bound, the Federal Reserve engaged
in the unconventional policy of large scale asset purchases (i.e. quantitative easing (QE)), with the
first announcement coming in late 2008. For FOMC meetings that involved announcements about
QE it is plausible that these announcements did not affect the expected fed funds rate over the next
two years (and thus the fed funds futures markets). Moreover it is reasonable to expect that the
stock market would react to announcements that reveal information about QE.

To check whether our results are driven by this period, we rerun our estimation excluding the
zero lower bound episode. The first two columns of table 2.8 present these results. Column 1a
shows that the overall response to exogenous monetary policy shocks and Delphic shocks is similar
to the baseline case reported in table 2.7, with similar standard errors as well. The interaction terms
with the unscheduled and turning point FOMC meetings also paint a similar picture. Relative to
the baseline results, on these particular FOMC meetings, the stock response to exogenous monetary
policy shocks is slightly more negative and the response to Delphic shocks is slightly less positive.12
Both the interaction terms are significant with p-values of 0.05 and 0.05 respectively.

Next we focus on the FOMC meetings in the early 1990s. Starting with February 1994, the
FOMC started releasing a statement to accompany its monetary policy decision. To check if
our results are driven by the 1991-1993 sample, we rerun the second step regressions using data
starting with the February 1994 FOMC meeting. Columns 2a and 2b report these results. The
12 We also tried truncating the sample in late 2007 to coincide with the beginning of turmoil in

the financial markets. The results are very similar to the ones presented here.

73

overall response to both exogenous monetary policy and Delphic shocks is slightly larger for the
post-1994 sample. For the interaction coefficients we find that the sign of the responses is similar to
the baseline case. The magnitude of the effects is a little larger for the exogenous monetary policy
shock and a little smaller for the Delphic shock on these particular FOMC meetings. However, the
standard errors are somewhat larger in this case.

Finally, we control for the employment report when running our regressions. Recall that the un-
derlying identifying assumption is that no other important macroeconomic event or announcement
is occurring in the relevant window around the FOMC announcement. However, as pointed out by
Gürkaynak, Sack, and Swanson (2005) there are a handful of FOMC meetings that coincide with
macro news releases. Specifically, in the early 1990s there are 7 FOMC meetings that occur on
the same day as the release of the employment report. Of special concern are 5 of these meetings
that are unscheduled because if the Federal Reserve and the stock market are both responding
to the employment report then our estimates will be mistakenly picking up that relationship. As
discussed above, in constructing the stock price change and monetary policy surprises the narrow
30 minute window was preferred precisely to avoid this particular issue. Gürkaynak, Sack, and
Swanson (2005) show that using the narrow 30 minute window does indeed help in circumventing
this identification issue. Here we confirm that our main results are not affected by excluding the 7
FOMC meetings that coincide with the employment report. Column 3a of table 2.8 shows that the
coefficients on the exogenous monetary policy and Delphic shocks are very similar to the baseline
results in table 2.7. Column 3b shows that on the unscheduled and turning point FOMC meetings,
the stock price response is in the same direction as the baseline results with the p-value on the
interaction term for the exogenous monetary policy shock and the Delphic shock being 0.04 and
0.03 respectively. Excluding the employment report in fact makes the magnitude of these effects a
little larger.

Overall, we conclude that our results are robust to sample selection. Next we use a VAR based

decomposition to further understand the stock price response.

74

2.4.4 VAR Based Decomposition

Here we try to understand in more depth the reason behind the observed reaction of the stock market
to monetary policy found in the previous section. In section 2.2.2 we discussed a broad but abstract
framework where stock price movements can be broadly attributed to two main components: i)
news about discount rates and ii) news about dividends (or cash flow news). In this section we use
a more concrete decomposition of stock prices based on the work of Campbell and Shiller (1988)
which calculates how much of the excess stock return can be attributed to expectations of future
interest rates, excess returns and dividends. In this framework we can evaluate if the Delphic shock
differentially affects the three component of the total excess stock return relative to the exogenous
shock. Additionally we can investigate the decomposition effects of the asymmetry on turning
point and unscheduled FOMC meetings.

The exact methodology used here follows the work of Bernanke and Kuttner (2005) and
Campbell and Ammer (1993). The key idea is to decompose the current period’s unexpected
t+1), future
excess returns (ey

t+1) into revisions of expectations of discounted future dividends ((cid:101)ed
t+1) and the real interest rate ((cid:101)er
t+1 −(cid:101)er
t+1 =(cid:101)ed

excess returns ((cid:101)ey

t+1)13

t+1

ey

where

(2.13)

(2.14)

t+1 −(cid:101)ey
∞

∞

∞


j=0

j=0

j=1

t+1 = (Et+1 − Et)

(cid:101)ed
(cid:101)er
(cid:101)ey
t+1 = (Et+1 − Et)

t+1 = (Et+1 − Et)

ρj ∆dt+1+j

ρj ∆rt+1+j

ρj ∆yt+1+j

ρ is the steady state level of the price to dividend ratio and is set to .9962 following BK. The
expectations terms in 2.14 need to be estimated to evaluate the decomposition in equation 2.13. A
13The details of the derivation can be found in Bernanke and Kuttner (2005) and Campbell and

Ammer (1993).

75

vector autoregression is used to construct these expectations. Campbell and Ammer (1993) show
how this relationship can be modeled using the variables of interest and any other variables that
might be helpful in forecasting excess returns. The resulting model is a six variable VAR with one
lag.

zt = Azt−1 + wt

(2.15)

The endogenous variables (zt) include the excess stock return, real interest rate, relative 3-month
T-bill rate, change in the 3-month T-bill rate, the dividend-price ratio, and the spread between the
10-year and 1-month Treasury yields. From this VAR we can estimate the variables of interest in
equation 2.13 using the following equations

(cid:101)ey
ey
t+1 = sywt+1
t+1 = sy ρA(I − ρA)−1
(cid:101)er
t+1 = sr (I − ρA)−1
t+1 +(cid:101)ey
t+1 +(cid:101)er
(cid:101)ed
wt+1
t+1 = ey

t+1

wt+1

(2.16)

where sy and sr are vectors with zeros and ones to pick out the relevant variables. The variance
of the current excess equity return can be decomposed into the sum of the three variances and
covariances.

t+1) = Var((cid:101)ed
t+1) + Var((cid:101)er
t+1) + Var((cid:101)ey
t+1)
−2Cov((cid:101)ed
t+1) − 2Cov((cid:101)ed
t+1,(cid:101)er
t+1) + 2Cov((cid:101)ey
t+1,(cid:101)ey
t+1,(cid:101)er
t+1)

Var(ey

(2.17)

Using monthly data from 1991 to 2011 (to match our baseline estimation sample), we report the
variance decomposition of excess equity returns in table 2.9. For ease of comparison, the first two
columns present the results from BK where they use data from 1989 to 2002. The left column
shows the total contribution to the variance while the second column shows the shares (divided by
Var(ey
t+1)). The majority of variation in excess returns is accounted for by the variance in expected
dividends and expected future excess returns. Relative to the BK results, our data suggest a slightly
bigger role for dividends (42% vs. 32%) and a smaller role for future excess returns (28% vs.

76

38%). At this stage, we should mention that there is recent work that points out some potential
issues with this framework. These concerns are primarily related to the residual based nature of
the decomposition, see for example the work of Chen and Zhao (2009). Thus we do not want
to place too much emphasis on how our results compare to BK because of the differences in the
sample dates and how the monetary policy surprises are constructed. Rather, the main purpose
of the analysis in this section is to compare how the decomposition varies between the exogenous
and Delphic shocks. We can more reasonably expect that the shortcomings of this residual based
decomposition are not systematically related to the manner in which we construct the exogenous
and Delphic shocks. Thus our emphasis will be on the difference in the decomposition between the
exogenous shock and Delphic shock rather than on the level of the effects themselves.

In this framework, a natural way to evaluate the effect of monetary policy is to include the

exogenous and Delphic shock directly in the VAR. Denoting the estimated exogenous shock by(cid:98)et
and the estimated Delpic shock(cid:104)(cid:99)γi, j

(cid:16)(cid:98)ΩCB
(cid:17)(cid:105) by(cid:101)gt we get
i,t+j|t −(cid:98)ΩM
zt = Azt−1 + φ1(cid:98)et + φ2(cid:101)gt +(cid:101)wt

i,t+j|t

(2.18)

The VAR is estimated at a monthly frequency which requires aggregating the monetary policy
shocks from the FOMC meeting frequency to a monthly frequency. We follow a simple rule of
summing up any monetary policy shocks in a given month to get the monthly number. We have
also tried aggregating the monetary policy shocks following the methodology of Gertler and Karadi
(2015a). The results from this alternative aggregation procedure are very similar to our baseline
case and are presented in the online appendix.

Having estimated the VAR, we want to calculate the effect of the two monetary policy shocks
on the discounted sums in equation 2.14. We can use the relationship outlined above in equation
2.16 together with the orthogonality of the monetary policy shocks. For example, consider the
equation for the real interest rate

(cid:101)er
t+1 = sr (I − ρA)−1

wt+1 = sr (I − ρA)−1 (φ1(cid:98)et + φ2(cid:101)gt +(cid:101)wt)

(2.19)

From this equation the effect of the exogenous shock on the present value of current and expected

77

future real rates is given by

sr (I − ρA)−1

φ1

(2.20)

and the effect of the Delphic shock on the present value of current and expected future real rates is
given by

sr (I − ρA)−1

φ2

(2.21)

The response of the present value of current and expected future excess returns and dividends
is calculated in a similar way. To account for the parameter uncertainty of the VAR coefficients in
A, standard errors are calculated using the delta method following Campbell and Ammer (1993)
and Bernanke and Kuttner (2005). Table 2.10 shows the response of the discounted sums to i)
the composite monetary policy surprise, ii) the exogenous shock and iii) the Delphic shock. For
ease of comparison we reproduce the results from BK in the first column where the sample runs
from 1989 to 2002. In the next 3 columns we present the results where both the VAR and the
monetary policy shocks are estimated using the 1991 to 2011 sample. For the second column we
replace the exogenous and Delphic shocks with the composite monetary policy surprise in the VAR
(equation 2.18). Relative to BK, the monetary policy surprise has a larger effect on current excess
equity return. Note this is not surprising as our monetary policy surprise measure contains forward
guidance surprises in addition to the federal funds rate surprises used in BK. However as found
in BK, the current excess return is explained mostly by discounted sums of dividends and future
excess returns.14

Relative to the composite monetary policy surprise, the exogenous shock (shown in the third
column) has a very similar effect on current excess returns. The size of the impact is slightly larger
(-17.5 vs. -16.7), which is consistent with the regressions from section 2.4.3. This larger negative
response is driven mostly by a larger positive response of future excess returns (4.0 vs 3.6). The
response to the Delphic shock are quite different, although the standard errors are substantially
larger. The overall effect on current excess returns is smaller at -12.0. The most interesting aspect
14In recent work Maio (2013) and Weber (2015) similarly find that the effect of monetary policy

is primarily driven by the response of expected future dividends.

78

is the composition of this response. The share of the dividend response is much bigger at -9.7,
accounting for 81% of the total effect on current excess returns (relative to 66% for the exogenous
shock).

Next we extend the above analysis to account for the differential effects on unscheduled and
turning point FOMC meetings. This can be done in a straightforward manner using the framework
of equation 2.18. Denote the unscheduled and turning point dummy by Dt.

zt = Azt−1 +(cid:101)φ1(cid:98)et +(cid:101)φ2(cid:101)gt +(cid:101)φ3Dt +(cid:101)φ4(cid:98)et Dt +(cid:102)φ5(cid:101)gt Dt +(cid:101)wt

(2.22)

(2.23)

(2.24)

Using this equation the effect on the various components can be calculated as above. For example,
on unscheduled and turning point FOMC meetings the effect of the exogenous shock on the present
value of current and expected future real rates is given by

and the effect of the Delphic shock is given by

sr (I − ρA)−1(cid:16)(cid:101)φ1 +(cid:101)φ4
sr (I − ρA)−1(cid:16)(cid:101)φ2 +(cid:101)φ5

(cid:17)
(cid:17)

Table 2.11 shows theses estimates. The response of current excess returns and its components to the

exogenous shock ((cid:101)φ1) is similar to that reported in table 2.10. The interaction effects of exogenous
shocks ((cid:101)φ4) are small as well. The overall response of current excess returns to a Delphic shock is
FOMC meetings. Specifically the total effect on these meetings ((cid:101)φ2 +(cid:101)φ5 = 13.5) is roughly the

more negative once we allow for the interaction (-21.1 vs. -12.0). This larger negative response on
regular FOMC days is counteracted by a large positive response on unscheduled and turning point

same size as the baseline effect from table 2.10 but with the opposite sign. This positive response is
mainly driven by a large fall in the future excess return and to a lesser extent by a rise in dividends
in response to contractionary Delphic shocks. The VAR decomposition exercise confirms that the
stock market responds very differently to Delphic shocks that occur on unscheduled or turning
point FOMC meetings. Moreover, the results point to a change in the risk premium as a major

79

In recent work Hanson and Stein (2015) and Gertler and
driver of this asymmetric response.15
Karadi (2015a) find that monetary policy shocks have substantial effects on bond interest rate term
premia. Our results show that, at least on certain FOMC dates, the stock risk premium also seems
to respond to monetary policy shocks. We view our results as providing complementary evidence
to this active area of research.

2.5 Conclusion

What are the effects of monetary policy on the economy? In this paper we aim to shed light on the
relatively unexplored information (or signalling) channel of the monetary transmission mechanism.
We conduct our analysis using the reaction of the stock market as a laboratory. By exploiting
differences in central bank and private sector forecasts we construct a measure of Federal Reserve
private information. We use this measure to separate monetary policy surprises into exogenous and
Delphic shocks. Exogenous shocks are surprise changes in monetary policy which are unrelated to
macroeconomic fundamentals whereas Delphic shocks are surprise changes in policy attributable
to the Federal Reserve’s private information about the state of the economy.

We find that, on average, stock prices fall more in response to contractionary exogenous
shocks relative to Delphic ones. However, on unscheduled and turning point FOMC meetings,
contractionary Delphic shocks actually result in an increase in stock prices. The results highlight
an unconventional channel of monetary transmission where contractionary policy actions can
stimulate the economy. An additional important implication of our results is that an FOMC that
is concerned with financial market reaction should pay extra attention to its statements on turning
point and unscheduled meetings.

A promising possibility for future work includes analyzing firm and industry level responses to
the exogenous monetary policy and Delphic shocks. Heterogeneous firm-level responses may be
informative about which kind of firms or industries are particularly sensitive to the revelation of
Federal Reserve private information.

15While the response of future excess returns is precisely estimated (p-value < 0.01), standard
errors in general are somewhat large and thus these results should be interpreted with some caution.

80

APPENDIX

81

Table 2.1: FOMC, Greenbook & Blue Chip Dates

FOMC

Greenbook Blue Chip Unsched/TP

X

X

X

X

X

X

X

X
X

Greenbook Blue Chip Unsched/TP
FOMC
12-Dec-90
8-Jan-91
30-Jan-91
1-Feb-91
30-Jan-91
7-Feb-91
30-Jan-91
8-Mar-91
20-Mar-91
27-Mar-91
20-Mar-91
30-Apr-91
8-May-91
15-May-91
28-Jun-91
5-Jul-91
28-Jun-91
6-Aug-91
14-Aug-91
21-Aug-91
14-Aug-91
13-Sep-91
25-Sep-91
2-Oct-91
25-Sep-91
30-Oct-91
30-Oct-91
6-Nov-91
30-Oct-91
6-Dec-91
11-Dec-91
18-Dec-91
11-Dec-91
20-Dec-91
30-Jan-92
6-Feb-92
25-Mar-92
1-Apr-92
9-Apr-92
25-Mar-92
20-May-92 14-May-92
26-Jun-92
2-Jul-92
19-Aug-92
13-Aug-92
13-Aug-92
4-Sep-92
30-Sep-92
7-Oct-92
12-Nov-92
18-Nov-92
16-Dec-92
23-Dec-92
29-Jan-93
4-Feb-93
24-Mar-93
17-Mar-93
19-May-93 14-May-93
30-Jun-93
8-Jul-93
11-Aug-93
18-Aug-93
15-Sep-93
22-Sep-93
10-Nov-93
17-Nov-93
22-Dec-93
15-Dec-93
31-Jan-94
4-Feb-94
16-Mar-94
22-Mar-94
18-Apr-94
16-Mar-94
17-May-94 13-May-94
30-Jun-94
6-Jul-94
16-Aug-94
12-Aug-94
21-Sep-94
27-Sep-94
9-Nov-94
15-Nov-94
14-Dec-94
20-Dec-94
25-Jan-95
1-Feb-95
28-Mar-95
22-Mar-95

10-Dec-90
10-Jan-91
10-Jan-91
10-Jan-91
10-Mar-91
10-Mar-91
10-May-91
10-Jun-91
10-Jun-91
10-Aug-91
10-Aug-91
10-Sep-91
10-Sep-91
10-Oct-91
10-Oct-91
10-Dec-91
10-Dec-91
10-Jan-92
10-Mar-92
10-Mar-92
10-May-92
10-Jun-92
10-Aug-92
10-Aug-92
10-Sep-92
10-Nov-92
10-Dec-92
10-Jan-93
10-Mar-93
10-May-93
10-Jun-93
10-Aug-93
10-Sep-93
10-Nov-93
10-Dec-93
10-Jan-94
10-Mar-94
10-Mar-94
10-May-94
10-Jun-94
10-Aug-94
10-Sep-94
10-Nov-94
10-Dec-94
10-Jan-95
10-Mar-95

TP

X

X

X

26-Jun-96
15-Aug-96
18-Sep-96
6-Nov-96
12-Dec-96
29-Jan-97
19-Mar-97

25-Jun-97
14-Aug-97
24-Sep-97
6-Nov-97
11-Dec-97
28-Jan-98
25-Mar-98

28-Jun-95
16-Aug-95
20-Sep-95
8-Nov-95
14-Dec-95
26-Jan-96
21-Mar-96

23-May-95 17-May-95 10-May-95
10-Jun-95
6-Jul-95
10-Aug-95
22-Aug-95
10-Sep-95
26-Sep-95
10-Nov-95
15-Nov-95
19-Dec-95
10-Dec-95
10-Jan-96
31-Jan-96
26-Mar-96
10-Mar-96
21-May-96 16-May-96 10-May-96
10-Jun-96
3-Jul-96
10-Aug-96
20-Aug-96
24-Sep-96
10-Sep-96
10-Nov-96
13-Nov-96
10-Dec-96
17-Dec-96
10-Jan-97
5-Feb-97
25-Mar-97
10-Mar-97
20-May-97 15-May-97 10-May-97
2-Jul-97
10-Jun-97
10-Aug-97
19-Aug-97
10-Sep-97
30-Sep-97
10-Nov-97
12-Nov-97
10-Dec-97
16-Dec-97
4-Feb-98
10-Jan-98
31-Mar-98
10-Mar-98
19-May-98 14-May-98 10-May-98
10-Jun-98
1-Jul-98
10-Aug-98
18-Aug-98
10-Sep-98
29-Sep-98
15-Oct-98
10-Sep-98
10-Nov-98
17-Nov-98
10-Dec-98
22-Dec-98
10-Jan-99
3-Feb-99
30-Mar-99
10-Mar-99
18-May-99 13-May-99 10-May-99
30-Jun-99
10-Jun-99
10-Aug-99
24-Aug-99
10-Sep-99
5-Oct-99
10-Nov-99
16-Nov-99
10-Dec-99
21-Dec-99
2-Feb-00
10-Jan-00
21-Mar-00
10-Mar-00
16-May-00 11-May-00 10-May-00
10-Jun-00
28-Jun-00
10-Aug-00
22-Aug-00
10-Sep-00
3-Oct-00
15-Nov-00
10-Nov-00

24-Jun-98
13-Aug-98
23-Sep-98
23-Sep-98
12-Nov-98
16-Dec-98
28-Jan-99
24-Mar-99

23-Jun-99
18-Aug-99
29-Sep-99
10-Nov-99
15-Dec-99
27-Jan-00
15-Mar-00

21-Jun-00
16-Aug-00
27-Sep-00
8-Nov-00

TP

TP

TP
X

TP

FOMC

19-Dec-00
3-Jan-01
31-Jan-01
20-Mar-01
18-Apr-01
15-May-01
27-Jun-01
21-Aug-01
2-Oct-01
6-Nov-01
11-Dec-01
30-Jan-02
19-Mar-02
7-May-02
26-Jun-02
13-Aug-02
24-Sep-02
6-Nov-02
10-Dec-02
29-Jan-03
18-Mar-03
6-May-03
25-Jun-03
12-Aug-03
16-Sep-03
28-Oct-03
9-Dec-03
28-Jan-04
16-Mar-04
4-May-04
30-Jun-04
10-Aug-04
21-Sep-04
10-Nov-04
14-Dec-04
2-Feb-05
22-Mar-05
3-May-05
30-Jun-05
9-Aug-05
20-Sep-05
1-Nov-05
13-Dec-05
31-Jan-06
28-Mar-06
10-May-06

X

X/TP

Greenbook Blue Chip Unsched/TP
13-Dec-00
13-Dec-00
25-Jan-01
14-Mar-01
14-Mar-01
9-May-01
20-Jun-01
16-Aug-01
27-Sep-01
31-Oct-01
5-Dec-01
23-Jan-02
13-Mar-02
1-May-02
20-Jun-02
7-Aug-02
18-Sep-02
30-Oct-02
4-Dec-02
22-Jan-03
13-Mar-03
30-Apr-03
18-Jun-03
6-Aug-03
10-Sep-03
22-Oct-03
3-Dec-03
21-Jan-04
11-Mar-04
28-Apr-04
23-Jun-04
5-Aug-04
15-Sep-04
3-Nov-04
8-Dec-04
26-Jan-05
16-Mar-05
28-Apr-05
22-Jun-05
4-Aug-05
14-Sep-05
26-Oct-05
7-Dec-05
25-Jan-06
22-Mar-06
3-May-06

10-Dec-00
10-Dec-00
10-Jan-01
10-Mar-01
10-Mar-01
10-May-01
10-Jun-01
10-Aug-01
10-Sep-01
10-Oct-01
10-Dec-01
10-Jan-02
10-Mar-02
10-Apr-02
10-Jun-02
10-Aug-02
10-Sep-02
10-Oct-02
10-Dec-02
10-Jan-03
10-Mar-03
10-Apr-03
10-Jun-03
10-Aug-03
10-Sep-03
10-Oct-03
10-Nov-03
10-Jan-04
10-Mar-04
10-Apr-04
10-Jun-04
10-Aug-04
10-Sep-04
10-Nov-04
10-Dec-04
10-Jan-05
10-Mar-05
10-Apr-05
10-Jun-05
10-Jul-05
10-Sep-05
10-Oct-05
10-Dec-05
10-Jan-06
10-Mar-06
10-May-06

TP

FOMC
29-Jun-06
8-Aug-06
20-Sep-06
25-Oct-06
12-Dec-06
31-Jan-07
21-Mar-07
9-May-07
28-Jun-07
7-Aug-07
18-Sep-07
31-Oct-07
11-Dec-07
22-Jan-08
30-Jan-08
18-Mar-08
30-Apr-08
25-Jun-08
5-Aug-08
16-Sep-08
8-Oct-08
29-Oct-08
16-Dec-08
28-Jan-09
29-Apr-09
24-Jun-09
12-Aug-09
23-Sep-09
4-Nov-09
16-Dec-09
27-Jan-10
16-Mar-10
28-Apr-10
23-Jun-10
10-Aug-10
21-Sep-10
3-Nov-10
14-Dec-10
26-Jan-11
15-Mar-11
27-Apr-11
22-Jun-11
9-Aug-11
21-Sep-11
2-Nov-11
13-Dec-11

X

X

TP

Greenbook Blue Chip Unsched/TP
21-Jun-06
3-Aug-06
13-Sep-06
18-Oct-06
6-Dec-06
24-Jan-07
14-Mar-07
2-May-07
20-Jun-07
2-Aug-07
12-Sep-07
24-Oct-07
5-Dec-07
5-Dec-07
23-Jan-08
13-Mar-08
23-Apr-08
18-Jun-08
30-Jul-08
10-Sep-08
10-Sep-08
22-Oct-08
10-Dec-08
22-Jan-09
22-Apr-09
17-Jun-09
6-Aug-09
16-Sep-09
29-Oct-09
9-Dec-09
20-Jan-10
10-Mar-10
21-Apr-10
16-Jun-10
4-Aug-10
15-Sep-10
27-Oct-10
8-Dec-10
19-Jan-11
9-Mar-11
20-Apr-11
15-Jun-11
3-Aug-11
14-Sep-11
26-Oct-11
7-Dec-11

10-Jun-06
10-Jul-06
10-Sep-06
10-Oct-06
10-Dec-06
10-Jan-07
10-Mar-07
10-Apr-07
10-Jun-07
10-Jul-07
10-Sep-07
10-Oct-07
10-Dec-07
10-Dec-07
10-Jan-08
10-Mar-08
10-Apr-08
10-Jun-08
10-Jul-08
10-Sep-08
10-Sep-08
10-Oct-08
10-Dec-08
10-Jan-09
10-Apr-09
10-Jun-09
10-Aug-09
10-Sep-09
10-Oct-09
10-Dec-09
10-Jan-10
10-Mar-10
10-Apr-10
10-Jun-10
10-Aug-10
10-Sep-10
10-Oct-10
10-Dec-10
10-Jan-11
10-Mar-11
10-Apr-11
10-Jun-11
10-Jul-11
10-Sep-11
10-Oct-11
10-Dec-11

This table reports all FOMC dates from 1991-2011 and the corresponding Greenbook and Blue Chip forecast dates used to construct
private information variables. Unscheduled FOMC decisions are denoted with an X and turning point decisions are denoted with TP.

82

Table 2.2: Summary Statistics - Monetary Policy Surprise and S&P 500

Monetary Policy Surprise
Mean
Median
Standard Deviation
Min
Max
Correlation
Observations
S&P 500 Return
Mean
Median
Standard Deviation
Min
Max
Correlation
Observations

All FOMC days
30 Minute Daily

Turning points

Unscheduled

30 Minute Daily

30 Minute Daily

0.00
0.01
0.09
-0.44
0.24

0.34
0.24
1.23
-2.94
5.14

0.00
0.01
0.07
-0.34
0.15

0.81
183

-0.24
-0.06
0.63
-1.88
4.08

0.47
183

-0.03
-0.08
0.12
-0.17
0.16

1.08
0.82
2.19
-2.27
5.01

-0.04
-0.03
0.13
-0.19
0.15

0.95
8

0.69
0.30
1.64
-0.75
4.08

0.93
8

-0.14
-0.14
0.15
-0.44
0.22

0.60
0.38
1.69
-1.13
5.01

-0.10
-0.10
0.12
-0.34
0.13

0.82
17

0.53
0.03
1.25
-0.92
4.08

0.90
17

This table reports the summary statistics calculated using a tight 30 minute window and a broad
daily window. The monetary policy surprise measure reported in percentage points is constructed
using a principal component analysis of futures data, see section 2.3.1 for details. The S&P 500
return is also reported in percentage points.

83

Table 2.3: Stock Market Response to Monetary Policy Surprise - 30 Minute Window

VARIABLES

MP Surprise

Unscheduled/Turning Point Dummy

MP Surprise x Unscheduled/Turning Point

Unscheduled FOMC Dummy

MP Surprise x Unscheduled

Turning Point Dummy

MP Surprise x Turning Point

Constant

S&P 500 (30 Minute Window)
(1)
(4)

(3)

(2)

-4.62
(0.84)

-4.34
(0.83)

-5.14
(0.93)

-4.24
(0.98)
0.05
(0.13)
-1.37
(1.60)

-0.01
(0.17)
-1.15
(1.74)

0.38
(0.28)
-4.70
(3.81)
-0.05
(0.03)

-0.02
(0.04)

-0.04
(0.04)

-0.03
(0.04)

Observations
R-squared
Adjusted R-squared

183
0.32
0.32

183
0.33
0.32

183
0.32
0.31

183
0.38
0.37

The table reports the regression of the change in the S&P 500 index on the monetary policy surprise,
both measured in a 30 minute window around FOMC announcements. The Unscheduled dummy
is set to 1 for FOMC meetings occurring outside the regularly scheduled dates. The Turning Point
dummy is set to 1 if the policy decision changed the fed funds rate in the opposite direction of the
previous change. The Unscheduled/Turning Point dummy is set to 1 for either occurrence. Robust
standard errors are in the parentheses.

84

Table 2.4: Stock Market Response to Monetary Policy Surprise - Daily Window

VARIABLES

MP Surprise

Unscheduled/Turning Point Dummy

MP Surprise x Unscheduled/Turning Point

Unscheduled FOMC Dummy

MP Surprise x Unscheduled

Turning Point Dummy

MP Surprise x Turning Point

Constant

S&P 500 (Daily Window)
(1)
(4)

(2)

(3)

-2.68
(2.10)

-1.57
(1.45)

-2.42
(1.45)

-1.55
(2.27)
-0.15
(0.34)
-2.21
(3.23)

-0.08
(0.47)
0.30
(3.25)

0.31
(0.42)
-12.16
(3.84)
0.31
(0.09)

0.34
(0.09)

0.33
(0.10)

0.35
(0.10)

Observations
R-squared
Adjusted R-squared

183
0.03
0.02

183
0.04
0.02

183
0.03
0.02

183
0.10
0.08

The table reports the regression of the change in the S&P 500 index on the monetary policy surprise,
both measured using a daily window around FOMC announcements. The Unscheduled dummy is
set to 1 for FOMC meetings occurring outside the regularly scheduled dates. The Turning Point
dummy is set to 1 if the policy decision changed the fed funds rate in the opposite direction of the
previous change. The Unscheduled/Turning Point dummy is set to 1 for either occurrence. Robust
standard errors are in the parentheses.

85

Table 2.5: 1st Step - Monetary Policy Surprise on Private Information

VARIABLES

MP Surprise

CPI0Q

U0Q

GDP0Q

IP0Q

CPI4Q

U4Q

GDP4Q

IP4Q

CPI0Q Lag

U0Q Lag

GDP0Q Lag

IP0Q Lag

CPI4Q Lag

U4Q Lag

GDP4Q Lag

IP4Q Lag

Constant

Observations
R-squared
Adjusted R-squared
P-Value

0.008
(0.004)
-0.005
(0.051)
0.018
(0.009)
0.003
(0.003)
0.013
(0.020)
0.025
(0.028)
0.014
(0.016)
0.006
(0.011)
-0.003
(0.007)
0.030
(0.044)
-0.007
(0.009)
0.002
(0.003)
0.002
(0.020)
-0.048
(0.027)
0.014
(0.017)
-0.018
(0.010)
0.012
(0.009)

182
0.16
0.08
0.01

The table reports the regression of the monetary policy surprise on the private information variables (con-
structed as the difference between the Greenbook forecasts and Blue Chip forecasts). “0Q" and “4Q" refer
to the nowcast and 4 quarter ahead forecast, see the main text for more details. The p-value is for the test of
joint significance of all the private info variables in the regression. Robust standard errors are in parentheses.

86

Table 2.6: Summary Statistics - Exogenous and Delphic Shocks

All Unscheduled/TP Pre-ZLB Post-1994 No Report

Exogenous Shock
Mean
Median
Standard Deviation
Min
Max
Correlation with MP surprise
Observations
Delphic Shock
Mean
Median
Standard Deviation
Min
Max
Correlation with MP surprise
Observations

0.00
0.01
0.06
-0.30
0.16
0.92
182

0.00
0.00
0.03
-0.09
0.08
0.40
182

-0.07
-0.04
0.11
-0.30
0.11
0.98
23

-0.01
-0.01
0.03
-0.06
0.05
0.61
23

0.00
0.01
0.06
-0.30
0.16
0.92
159

0.00
0.00
0.03
-0.09
0.08
0.40
159

0.00
0.00
0.06
-0.30
0.16
0.90
139

0.00
0.01
0.03
-0.09
0.08
0.41
139

0.00
0.01
0.06
-0.30
0.16
0.91
175

0.00
0.01
0.03
-0.09
0.08
0.39
175

This table reports the summary statistics calculated using a tight 30 minute window. Both shocks,
reported in percentage points, are retrieved from the regression of monetary policy surprises on Fed
private information. The exogenous monetary policy (MP) shock is the residual and the Delphic
shock is the fitted value, see section 2.4.2 for details. The first column includes all FOMC dates
in our sample, the second includes only unscheduled and turning point dates, the third includes
all dates prior to the fed funds rate hitting the zero lower bound, the fourth column includes all
dates following 1994, and the fifth includes all dates that did not coincide with the release of an
unemployment report.

87

Table 2.7: 2nd Step - Response of Stock Prices to Exogenous and Delphic Shocks

VARIABLES

Exogenous Shock

Delphic Shock

Unscheduled/Turning Point Dummy

Exogenous x Unscheduled/Turning Point

Delphic x Unscheduled/Turning Point

Unscheduled FOMC Dummy

Exogenous x Unscheduled

Delphic x Unscheduled

Turning Point Dummy

Exogenous x Turning Point

Delphic x Turning Point

Constant

S&P 500 (30 minute window)
(1)

(2)

(3)

(4)

-5.19
(0.86)
-2.49
(1.80)

-4.74
(0.84)
-2.13
(1.91)

-5.72
(0.99)
-2.12
(1.87)

-4.62
(0.90)
-2.93
(1.97)
-0.01
(0.18)
-3.59
(2.16)
14.62
(7.08)

0.15
(0.29)
-2.48
(2.84)
18.4
(14.85)

0.06
(1.59)
-8.67
(12.57)
10.47
(59.07)
-0.05
(0.04)

-0.03
(0.04)

-0.04
(0.04)

-0.03
(0.04)

Observations
R-squared
Adjusted R-squared

182
0.34
0.33

182
0.39
0.37

182
0.37
0.35

182
0.42
0.40

The table reports the regression of the change in the S&P 500 index on the residual and fitted value
of the policy surprise from the first step, both measured in a 30 minute window around FOMC
announcements. The Unscheduled dummy is set to 1 for FOMC meetings occurring outside the
regularly scheduled dates. The Turning Point dummy is set to 1 if the policy decision changed the
fed funds rate in the opposite direction of the previous change. The Unscheduled/Turning Point
dummy is set to 1 for either occurrence. Bootstrapped standard errors are in the parentheses.

88

Table 2.8: 2nd Step - Subsamples

VARIABLES

Exogenous Shock

Delphic Shock

Unscheduled/Turning Point Dummy

Exogenous x Unscheduled/Turning Point

Delphic x Unscheduled/Turning Point

Constant

Observations
R-squared
Adjusted R-squared

Pre-ZLB

(1a)

(1b)

Post-1994
(2b)

(2a)

No Employment Report
(3a)

(3b)

-5.87
(1.09)
-1.27
(1.83)

-0.04
(0.05)

159
0.35
0.34

-4.11
(1.03)
-1.71
(1.97)
-0.03
(0.18)
-4.70
(2.34)
11.83
(6.00)
-0.07
(0.04)

159
0.41
0.39

-6.87
(1.17)
-2.24
(1.98)

-0.04
(0.05)

148
0.38
0.37

-5.25
(0.93)
-2.33
(2.09)
-0.11
(0.44)
-4.46
(4.45)
8.67
(11.78)
-0.06
(0.05)

148
0.41
0.39

-5.74
(1.05)
-2.31
(1.87)

-0.03
(0.04)

175
0.34
0.33

-4.56
(0.90)
-3.14
(2.01)
-0.13
(0.24)
-5.10
(2.54)
20.85
(9.61)
-0.04
(0.04)

175
0.41
0.39

The table reports the regression of the change in the S&P 500 index on the residual and fitted value
of the policy surprise from the first step, both measured in a 30 minute window around FOMC
announcements. The Unscheduled dummy is set to 1 for FOMC meetings occurring outside the
regularly scheduled dates. The Turning Point dummy is set to 1 if the policy decision changed the
fed funds rate in the opposite direction of the previous change. The Unscheduled/Turning Point
dummy is set to 1 for either occurrence. Bootstrapped standard errors are in the parentheses.

89

Table 2.9: Excess Stock Return Variance Decomposition

Var(Excess Return)

Var(Dividends)

Var(Real Rate)

Var(Future Returns)

-2*Cov(Dividends, Real Rate)

-2*Cov(Dividends, Future Excess Returns)

2*Cov(Future Excess Returns, Real Rate)

BK 1989 - 2002
Total Share (%)

1991 - 2011

Total Share (%)

19.00

6.10

0.10

7.20

-0.60

7.20

1.00

19.59

8.31

0.29

5.45

0.55

4.85

0.13

31.90

0.60

38.00

-3.20

37.70

5.10

42.43

1.50

27.82

2.80

24.77

0.68

The table reports the variance decomposition of current excess equity returns into the variances of
revisions in expectations of dividends, real interest rates, future excess returns, and the covariances
between them.

90

Table 2.10: Response of Excess Returns to Exogenous & Delphic Shocks

BK 1989 - 2002 MP Surprise Exog Shock Delphic Shock

Current Excess Ret.

Future Excess Ret.

Real Interest Rate

Dividends

-11.01
(3.72)
3.29
(1.10)
0.77
(1.87)
-6.96
(2.35)

-16.65
(5.09)
3.61
(2.75)
1.72
(0.63)
-11.32
(4.93)

-17.45
(5.47)
4.02
(2.94)
1.82
(0.68)
-11.60
(5.42)

-12.07
(12.52)
1.24
(4.63)
1.10
(1.54)
-9.73
(11.47)

This table reports the response of current excess equity returns and its components to monetary
policy shocks. The first column reproduces the BK results estimated on the sample 5/1989 to
12/2002. The remaining three columns use the baseline data sample of 2/1991 to 12/2011. Delta
method standard errors are in parentheses.

91

Table 2.11: Response of Excess Returns on Unscheduled/Turning Point Meetings

Exog Delphic Unsch/TP Dum Exog x Unsch/TP Delphic x Unsch/TP

(cid:101)φ5

(cid:101)φ1

(cid:101)φ2

(cid:101)φ3

(cid:101)φ4

Current Excess Ret.

Future Excess Ret.

Real Interest Rate

Dividends

-19.05
(6.13)
5.35
(2.98)
1.17
(0.74)
-12.52
(5.98)

-21.13
(14.19)
9.66
(5.15)
0.44
(1.73)
-11.03
(13.00)

0.35
(1.09)
-0.01
(0.34)
0.12
(0.13)
0.45
(1.01)

3.61
(14.02)
-1.89
(4.47)
3.21
(1.65)
4.93
(13.06)

34.67
(27.94)
-31.14
(8.72)
1.70
(3.29)
5.23
(25.83)

This table reports the response of current excess equity returns and its components to monetary
policy shocks interacted with the unscheduled/turning point dummy. The dummy equals 1 on dates
for which the FOMC decision was unscheduled or reversed the previous direction of policy. Delta
method standard errors are in parentheses.

92

Figure 2.1: Monetary Policy Surprise

93

91929394959697989900010203040506070809101112-0.4-0.3-0.2-0.100.10.230 minute window91929394959697989900010203040506070809101112-0.5-0.4-0.3-0.2-0.100.10.20.3Daily windowFigure 2.2: Private Information Variables

94

UnemploymentI PCPIGDP95000510Nowcast-2-101950005101 Qtr-202950005102 Qtr-202950005103 Qtr-101950005104 Qtr-10195000510-4-20295000510-202495000510-10195000510-1-0.500.595000510-1-0.5095000510-4095000510-40495000510-50595000510-4-20295000510-20295000510-0.200.20.495000510-0.400.495000510-0.500.595000510-0.500.5195000510-0.500.51Figure 2.3: Stock Returns vs Monetary Policy Surprises

95

MonetaryPolicySurprise30minwindow-0.4-0.200.2S&P500return30minwindow-2-101234512/20/199104/09/199202/04/199407/06/199510/16/199805/18/199906/30/199901/01/200104/18/200109/18/200712/11/200701/22/200810/29/200812/16/2008Turning PointsUnscheduledFigure 2.4: Exogenous and Delphic Shocks

96

91929394959697989900010203040506070809101112-0.3-0.2-0.100.1Exogenous Shock91929394959697989900010203040506070809101112-0.1-0.0500.050.1Delphic ShockCHAPTER 3

REGIONAL RESPONSES TO MONETARY POLICY OVER TIME

3.1 Introduction

The US is a large country with significant geographical diversity. This diversity includes
economic activities and conditions. Accounting for and understanding these differences is important
for policymakers looking to optimize policy at the national level. Since monetary policy in particular
is applied in a one-size-fits-all approach, understanding how each region responds to monetary
policy is a significant consideration.

Carlino and DeFina (1998) wrote the seminal paper on differential regional responses to mon-
etary policy shocks in the United States. Their analysis showed that real personal income in five of
the eight BEA regions do not have a statistically different response to policy shocks than aggregate
US real personal income does. On the other hand, three regions were shown to have significantly
different responses. The sample used in Carlino and DeFina (1998) ended in 1992, highlighting
a need to update their results. Additionally, evidence that the effects of monetary policy have
changed over time, as presented in Boivin, Kiley, and Mishkin (2010) among others, makes this
task especially imperative. Beyond updating the data used, it is also possible to update the empirical
method.

Carlino and DeFina (1998) estimate a ten variable VAR through recursive identification. Recent
papers such as Faust, Swanson, and Wright (2004b) have called into question the legitimacy of
VAR identification schemes which impose zero contemporaneous effect restrictions. Additionally,
the somewhat arbitrary ordering of the variables in a recursive VAR can have a significant effect on
the resulting impulse responses. These issues can both be overcome through external instrument
identification.

The role of various channels in transmitting monetary policy at the regional level is investigated
as well. Carlino and DeFina (1998) along with other papers in the literature have consistently found

97

evidence of a regional interest rate channel. Taking a similar approach, we analyze whether evidence
for this channel holds for a more recent period of time. Additionally we explore whether evidence
for a credit channel, broad or narrow, can be found. Further, we examine the relationship between
state-level effects of monetary policy and state-level industry shares of nonfarm employment.

3.2 Previous Literature

As noted in the introduction, Carlino and DeFina (1998) is the original paper in this literature.
Their sample period is 1958-1992 and they investigate the response of regional real personal
income to a monetary policy shock. Their regions are defined by the Bureau of Economic Analysis
categorization. The main finding is that five “core” regions – New England, Mideast, Plains,
Southeast, and Far West – respond to monetary policy shocks in a statistically equivalent manner
to the US aggregate response. Of the three non-core regions, the Great Lakes is significantly more
sensitive to a shock while the Southwest and Rocky Mountains are less sensitive.

A ten variable quarterly VAR is used including the fed funds rate, the eight regional measures of
real personal income growth, and a measure of the relative price of energy to control for aggregate
supply shocks. The same VAR specification is used here, with the only difference being the
identification approach. Carlino and DeFina (1998) also attempt to investigate the role of various
monetary policy channels in explaining the differential regional responses. They estimate 48 state-
level VARs with state real personal income replacing the regional measure. The state personal
income responses are regressed upon proxies for the interest rate channel, broad credit channel,
and narrow credit channel. The results provide strong evidence of a regional interest rate channel,
along with weaker evidence of a broad and narrow credit channel - although the coefficient on the
narrow credit channel is the opposite sign than expected.

Owyang and Wall (2003) perform a similar analysis using the BEA measures of regional personal
income. They estimate the responses of real regional personal income to a monetary policy shock
in a 12 variable VAR including the fed funds rate, CPI, the 10 year treasury rate, a commodity
price index, and the eight regional personal income measures, with recursive identification. Their

98

full sample results (1960-2002) show that the Great Lakes region suffers the largest total loss in
personal income following a contractionary monetary shock while New England, the Plains, and
the Mideast have the smallest responses.

They also split their sample into a pre-Volcker era (1960-1978) and a Volcker-Greenspan era
(1983-2002). Total personal income losses are larger in the pre-Volcker era compared to the full
sample. On the other hand, personal income losses are much smaller in the Volcker-Greenspan era,
with five regions actually seeing total personal income gains over the period. This is in line with
Boivin, Kiley, and Mishkin (2010) who find that monetary policy shocks have a weaker effect on
real activity at the national level post-1980. Schunk (2005) conducts a similar study to Owyang
and Wall (2003) and also reports that the magnitude and differences in state-level responses to
monetary shocks have decreased over time.

Owyang and Wall (2003) investigate the roles of the interest rate channel, broad credit channel,
and narrow credit channel in explaining the differential regional responses as well. They construct
19 sub-regions from the eight BEA regions and re-estimate their VAR at this sub-regional level.
Total personal income loss from each sub-region is regressed on proxies for the three monetary
channels as in Carlino and DeFina (1998). For the full sample, all three channels are found to
significantly impact the personal income losses resulting from a monetary tightening, although the
sign on the narrow credit channel is the opposite of what was expected (indicating regions with a
greater share of large banks experience greater personal income losses). For the Volcker-Greenspan
era only the interest rate channel is significant in determining personal income loss following a
monetary tightening.

Two other related studies are Kouparitsas (2001) and Crone (2005). Kouparitsas (2001) esti-
mates a ten variable VAR similar to Carlino and DeFina (1998), albeit with different identifying
restrictions, which includes real personal income from the eight BEA regions, the fed funds rate,
and oil prices for the sample 1969-2001. His results show the Great Lakes, Plains, and Rocky
Mountains to be the region’s most sensitive to monetary policy shocks while the Southwest is
reported as the least sensitive.

99

Crone (2005) estimates a VAR for the sample 1959-1993 that is identical to Carlino and DeFina
(1998) other than the definitions of regions used. He constructs a new set of regions based on state
business cycle patterns. His results show that his Great Lakes region (which is similar to the BEA
Great Lakes region) is the most sensitive to monetary policy shocks while what he calls the Energy
Belt (which is made up of portions of the BEA’s Southwest and Rocky Mountains regions) is the
least sensitive.

Beckworth (2010) estimates a state-level VAR which includes a national coincident indicator
(a measure of real economic activity published by the Philadelphia Fed), CPI, PPI, and 48 state
national coincident indicators for the sample 1983-2008. Using state coincident indicators rather
than state personal income allows for differential regional prices to be accounted for, since personal
income can only be deflated by national CPI. Overidentifying restrictions are used to surmount
the degrees of freedom problem. The results are quite similar to Crone (2005) as the states least
sensitive to monetary policy shocks tend to be from the so-called Energy Belt (mostly Southwest
and Rocky Mountains) while the most sensitive states tend to be from the so-called Rust Belt
(mostly Great Lakes).

Crone et al. (2007) provides a nice summary of the relevant studies: “the area around the Great
Lakes is one of the regions most affected by shocks to monetary policy. Regions with a large
proportion of their economic activity derived from energy are among the least affected, whether
this is the Southwest as in the traditional BEA definition of regions or the Energy Belt as I have
defined it.” Additionally, as far as investigations into the role of monetary policy channels have
gone, “differences in industry mix (interest rate channel) are the only explanation that has found
consistent support in economic studies of regional differences in the effects of monetary policy.”

3.3 Estimation

As the previous section discusses, other researchers have attempted updating Carlino and
DeFina (1998). The fact that these investigations have yielded relatively consistent results is
encouraging. Here we will attempt to improve on the past literature by re-estimating the baseline

100

VAR from Carlino and DeFina (1998) using the longest available sample (1958:Q1 – 2015:Q2) and
an external instruments identification strategy.

The previous studies in this literature all rely on recursive identification assumptions. As is well
known, impulse responses produced by recursive VARs can be sensitive to the somewhat arbitrary
choice of how the endogenous variables are ordered within the system. Additionally, the most
common recursive structure orders the federal funds rate last, which in the context of a quarterly
VAR has the interpretation that a shock to the fed funds rate will not impact the real variables
in the system until a full three months has passed. The validity of such an assumption has been
questioned in recent years in papers such as Faust, Swanson, and Wright (2004b).

The external instruments approach can alleviate the problems associated with arbitrary variable
ordering and zero contemporaneous effect assumptions. Developed by Stock and Watson (2012) and
Mertens and Ravn (2013), the approach has been applied in a similar monetary context by Gertler
and Karadi (2015b). Rather than relying on a recursive ordering of the endogenous variables, it
achieves identification by exploiting the information contained in a series of exogenous monetary
policy shocks. For a detailed, yet accessible explanation of the method see Mertens and Ravn
(2013). The measure of exogenous monetary policy shocks used here will be the Romer and Romer
(2004) (R&R) shock series. This series identifies monetary policy shocks as innovations to the fed
funds rate which are uncorrelated with the Greenbook forecasts made prior to each FOMC meeting
(see Romer and Romer (2004)).

We use the same ten variable VAR as Carlino and DeFina (1998) which includes: the effective
fed funds rate, real personal income growth in each of the eight BEA regions, and a measure of
relative energy prices (the PPI for fuels and related products relative to total PPI). The R&R shock
measure is used as an instrument to identify unanticipated innovations to the fed funds rate. We
first estimate cumulative impulse responses for the entire 226 quarter sample. Next, we split the
sample with the early period being 1958:Q1 – 1992:Q4 and the recent period being 1993:Q1 –
2015:Q2. This will be instructive for two reasons. First, comparing our early period results with
those of Carlino and DeFina (1998) will allow us to see whether any differences arise from using the

101

external instruments approach rather than a recursive one. Second, comparing our early period and
recent period results will allow us to see how regional responses to monetary policy have changed
over time.

3.4 Results

3.4.1 Full Sample

Figure 3.1 displays the cumulative impulse response functions for the full sample. Statistical
equivalence of each regional response with the national response is tested using a two sample t-test.
Five regions – the Far West, Plains, Rocky Mountains, Southeast, and Southwest – are found to
respond to monetary policy shocks in essentially the same manner as the nation as a whole. Three
regions – the Great Lakes, Mideast, and New England – are found to have significantly different
responses than the national. The Great Lakes response is the most severe of all the regions while the
Mideast and New England have the mildest responses. The F-statistic from the first stage regression
of the fed funds rate residual on the R&R shocks is 11.01, which is above the common cut off point
of 10 and therefore indicates that we needn’t be concerned with relevancy issues.

There are interesting comparisons here to Carlino and DeFina (1998). Three regions have
responses significantly different from the national in both, with the Great Lakes being the only
region to have a more severe response. The two regions having weaker responses are different
between the two sets of results, as it’s the Southwest and Rocky Mountains in Carlino and DeFina
(1998) versus the Mideast and New England here. It is interesting to note that the Rocky Mountains
have the second most severe response after 5 years in our results. The Southeast, Far West, and
Plains have responses equivalent to the national in both. The magnitude of personal income lost is
very similar for both sets of results as well. After five years national personal income loss is about
0.8% lower in Carlino and DeFina (1998) compared to 1.1% lower here, while the Great Lakes loss
is about 1.3% in each.

102

3.4.2 Early Sample

Figure 3.2 presents the cumulative impulse response functions for the early sample, 1958:Q2 –
1992:Q4. This is the sample originally estimated in Carlino and DeFina (1998). The F-statistic
here is 50.13, indicating that the R&R shocks are very strong instruments for this time period. Once
again, three regions have responses significantly different from the national with the Great Lakes
having the strongest. The Southeast response is also significantly stronger than the national in this
period while the Southwest is the only region to have a significantly weaker reaction.

These results are qualitatively similar to Carlino and DeFina (1998), as the Great Lakes has the
most sensitive response and the Southwest has the least sensitive response in each. The reactions
of the Plains, Mideast, New England, and Far West are equivalent to the national response in each
as well. The results here show a much larger magnitude of personal income lost however. The
total loss after 5 years for the Great Lakes region is 2.6% - which is twice as large as Carlino and
DeFina’s estimate. In fact, the cumulative loss is at least double for every region (and nationally)
except for New England, which increases from 0.8% in Carlino and DeFina (1998) to about 1.3%
here.

3.4.3 Recent Sample

Figure 3.3 displays the cumulative impulse response functions for the recent period, 1993:Q1 –
2015:Q2. The results here are novel, as no paper measuring regional responses to monetary policy
shocks has used a sample going past 2002. We now see a much different pattern in the responses.
All regions show an initial negative reaction to an unanticipated monetary tightening with personal
income loss hitting a trough after two quarters. After one year each region’s response begins
rebounding though, with a few even returning to positive gains at that point. After two years none
of the regions have a negative total personal income loss, although after five years there is a small
total loss for the Far West, Great Lakes, and Southwest.

The magnitude of real personal income lost or gained after five years is small, with the Far
West suffering the biggest loss of -0.45% and New England seeing the largest gain at 0.44%. The

103

national response is only negative for the first three quarters and after five years sees a slight gain of
about 0.35%. The only regions which have a statistically equivalent response are the Mideast and
New England. Importantly, none of the responses - regional or national - are statistically different
from zero. These results are in line with Owyang and Wall (2003) who report smaller effects of
monetary policy shocks on regional real personal income over time. This is also consistent with
Boivin, Kiley, and Mishkin (2010) who find smaller effects of monetary policy on national real
activity in the post-1984 period.

One concern with the results for this sample is the F-statistic of 9.16 from the first stage
regression, which is slightly below the standard cutoff of 10. This indicates we may have a
weak instrument problem. A possible cause of this low F-statistic is the unusual monetary policy
environment of the post-2008 United States, where the fed funds rate has been stuck at the zero
lower bound. We therefore re-estimate the recent sample from 1984-2008 as a robustness check.
These results are presented in Figure 3.4. The general pattern of the results is quite similar and once
again each region’s response is not statistically different from zero. The F-statistic here increases
to a safe 15.32, which reassures that our results are not distorted by a weak instrument.

3.5 Regional Monetary Channels

There is a consensus in the previous literature that manufacturing-intensive states are more
responsive to monetary policy shocks on average, indicating that there is an interest rate channel
of monetary policy at the regional level. Here we attempt a similar analysis using our more recent
dataset. While many channels of monetary policy have been proposed (see Boivin, Kiley, and
Mishkin (2010)) the papers in this literature have looked at just three: the interest rate channel, the
broad credit channel, and the narrow credit channel.

An interest rate channel works as follows: monetary tightening increases interest rates, which
increases the user cost of capital, reduces investment and thus reduces aggregate demand. As a
proxy for this channel we will use a state’s manufacturing share of total nonfarm employment, since
the manufacturing sector tends to be the most interest rate sensitive. A broad credit or balance

104

sheet channel states that monetary tightening should increase the cost of raising external funds, thus
decreasing investment and aggregate demand. Since smaller firms tend to face higher borrowing
costs than large entities they should be more adversely affected by a monetary tightening. We
therefore use a state’s share of total employment in firms with less than 100 employees as a proxy.
A narrow credit or bank lending channel states that a monetary tightening will decrease the supply
of loanable funds available from banks and thus firms reliant on bank credit will either be unable
to borrow or may only be able to borrow at a higher cost. This leads to a reduction in investment
spending and lowers aggregate demand. Small banks have less access to alternative sources of
funding than large banks do so they should be more intensely affected by a monetary tightening
working through this channel. Therefore as a proxy we will use a state’s share of total bank assets
in banks with less than $1 billion in assets.

Due to the consensus in the previous literature for earlier time periods and limitations on the
data available for our proxy variables we will only analyze the period 1990-2008 in this section.
This is of special interest because, as seen in the previous section, the response of regional personal
income to monetary policy shocks over this time has changed significantly from past decades. Thus
it will be of interest to see if there are changes in how monetary policy has been transmitted over
this period as well.

To achieve sufficient cross-sectional variation we run 48 state-level VARs (leaving out Alaska
and Hawaii). Each VAR will include the fed funds rate, the relative energy measure, state real
personal income growth, regional real personal income growth for that state’s region minus the
contribution of the state itself, and national real personal income growth minus the previous variable.
This is equivalent to the state-level VARs used in Carlino and DeFina (1998) except we use national
personal income growth rather than including personal income growth for each of the seven other
regions separately. Once again, we will implement external instrument identification with the R&R
shocks as our instrument for fed funds rate innovations.

The state-level cumulative impulse responses are similar to the regional responses analyzed in
the previous section, with most states hitting a trough in the second quarter following an unan-

105

ticipated monetary tightening. Therefore the dependent variable here will be the second quarter
cumulative decline in real personal income growth for each state. The three independent variables
will be the channel proxies discussed above. The variables are constructed as an average of the
annual observations available from 1990-2008 for each. State manufacturing share is available for
each of those years, small business share is available for 1992 and then every year of 1997-2008,
and small bank share is available for 2000, 2004, and 2008. We run the following regression with
our 48 observations:

PIresponsei = β1 + β2M ANUFi + β3SM ALLBUSi + β4SM ALLBANKi + 
i

(3.1)

The results are shown in table 3.1. We find that state share of small businesses is the only variable
to have any statistical significance, at the 10% level, providing evidence that a regional broad credit
channel may exist. The coefficient on small business share tells us that a doubling of a state’s share
of small firms is associated with an additional 1.5% loss in real personal income growth at the
trough following a surprise monetary tightening.

State manufacturing share and small bank share both appear to be insignificant for determining
losses to real personal income growth following a shock. Surprisingly, the coefficient on manufac-
turing share is the opposite of what would be expected. The past literature has found that states with
larger manufacturing sectors have significantly larger personal income losses following a mone-
tary tightening, indicating that the coefficient should be negative. The fact that it is positive and
insignificant here indicates that the interest rate channel may no longer be operative at the regional
level. The coefficient on the small bank share variable is very insignificant, both quantitatively
and statistically. This perhaps is not surprising given the increasing centralization of the financial
sector over time.

These results are interesting as they suggest that the change in regional responses to monetary
policy over time may be attributable to a change in how monetary policy is transmitted. In the

106

past, the interest rate sensitivity of a region’s economy is thought to have been the driving factor in
differential regional responses, while in the modern era this no longer seems to be the case

3.6 Regional Industry Composition

Another possible explanation as to why regional responses to monetary policy have diminished
over time is simply that differences in regional economies have diminished over time. As noted in
section 2, Crone et al. (2007) summary of the literature states that, “differences in industry mix are
the only explanation that has found consistent support in economic studies of regional differences
in the effects of monetary policy.” The previous literature, as in our previous section, has only
looked at differences in state manufacturing sectors however.

In this section we will investigate the relationship between state-level responses to positive
monetary policy shocks and state-level industry shares of total nonfarm employment for seven
NAICS/SIC industries. Industry codes changed in 1990 from the old SIC system to the current
NAICS codes. This matches up roughly with our two sample periods, and will give us enough
observations for the recent period to leave the post-2008 years out. Thus in this section our early
sample will be 1958-1989 and our recent sample will be 1990-2008.

We use the same 48 state-level VARs from section 5, this time including the early sample as
well. As in section 5, for the recent sample our dependent variable will be the second quarter
cumulative decline in real personal income for each state, since that is the trough of the response in
most cases. For the early sample the trough does not occur till the twentieth quarter for most states;
the decline in real personal income does begin to flatten out after the eighth quarter though, so we
will use the eighth quarter response as the dependent variable there.

Using data from the BLS state and metro area employment, hours, and earnings database we
construct average share of state nonfarm employment measures for each of the following industries
over the respective samples: mining and logging, construction, manufacturing, trade transport and
public utilities, financial activities, services, and government. Under the newer NAICS codes an
Including the
information industry exists that has no close analogue under the old SIC codes.

107

information sector in our recent sample regressions does little to change the results as it is never
significant, therefore we leave it out of our main specification to facilitate the comparison. We run
the following regression for each sample period:

PIresponsei = β1 + β2MIN INGi + β3CONST Ri + β4M ANUFi + β5TT PUi + β6FINi

(3.2)

+ β7SE RVi + β8GOVi + 
i

The results are presented in Table 3.2. In the early sample three industry shares are significantly
associated with the eighth quarter cumulative loss of real personal income. A more severe loss is
associated with a larger share of nonfarm employment in the manufacturing and service industries.
On the other hand, a weaker response is associated with a larger share of employment in the financial
sector. The coefficients can be interpreted as follows: a doubling of a state’s share of nonfarm
employment in, say manufacturing is associated with an additional 2.38% loss in real personal
income growth eight quarters after a positive monetary policy shock.

In the recent sample no industry shares are significantly associated with the real personal
income loss following a shock. Standard errors are quite large and the closest sector to being
statistically significant is mining and logging, however the p-value there is still just 0.18. These
results suggest that differential state responses to monetary policy can no longer be explained by
industry composition of the states.

One reason for this, as explored in section 5, is that monetary policy has been transmitted
to the real economy differently in the recent sample than in the early sample. Another possible
explanation however is that there is fewer differences in state-level industry composition in recent
decades. Table 3.3 shows the standard deviation of the average industry shares across 48 states
over the two samples. As can be seen, the variance in industry employment share of nonfarm labor
has declined in six of the seven industries, with finance being the only exception. This means that
states have become more similar to one another in industry composition over time.

108

Differences in the effects of monetary policy at the state level were, but no longer are associated
with differences in industry mix. Differences in industry mix have diminished over time as well.
A natural implication therefore is that the decrease in differential regional responses to monetary
policy over time has been driven by greater similarities in regional industry composition.

To determine which of these explanations holds more power - a change in the transmission
mechanism or a change in industry mix - it would be instructive to examine how differential
industry responses to monetary policy shocks have evolved over time. If industry responses have
not changed drastically over these two samples it would imply that the decrease in differential
regional responses is primarily due to a homogenization of state economies. On the other hand, if
industry responses have changed significantly it may imply that the monetary policy transmission
mechanism has changed over time – in which case such changes may be partially driving the
diminishing regional differences. This is an interesting avenue for further study.

3.7 Conclusion

Investigations into US regional responses to monetary policy shocks have produced an inter-
esting literature over the past two decades. This paper has updated the seminal contribution of
Carlino and DeFina (1998) by adding over 20 years of additional data and employing a cutting
edge identification strategy. Our results for the early sample are consistent with the past literature,
as we find real personal income growth in the Great Lakes region to respond most sensitively to
monetary policy shocks while real personal income growth in the Southwest region responds least
sensitively.

Our most interesting results are from the recent sample. The estimated responses show very
little reaction to a monetary policy shock for any region. Additionally, the differences among the
regional responses are negligible as none are statistically different from zero over 20 quarters. This
result is consistent with Boivin, Kiley, and Mishkin (2010) which finds that the impact of monetary
policy on real variables at the national level has been declining over time. Additionally, it supports
the argument from ? that choice of sample period is important for estimating regional responses to

109

monetary policy. Just looking at our full sample results could lead to very inaccurate conclusions
if one is concerned with the impact of monetary policy on regional personal income today.

Our results have a policy implication as well. Monetary policy shocks appear to have little to
no effect on regional real personal income in the present era. Thus Federal Reserve officials may
be able to disregard regional level effects when making policy decisions. It is unclear whether
officials have actually given weight to such regional effects in the past - nonetheless knowledge of
the existence of such differences was surely valuable. For the last two decades however it appears
that these differences, in real personal income at least, have largely vanished.

We explore the role of three different monetary policy channels in explaining the regional
responses. We find some evidence of a broad credit channel operating at the regional level. Just as
importantly, we find no evidence of an interest rate channel, which contradicts the results of past pa-
pers studying earlier periods. This suggests that the reduction in magnitudes and differences among
regional responses over time may be due to a change in the way monetary policy is transmitted to
the real economy. Alternatively, we also find evidence that the diminished regional differences may
be driven by a homogenization of industry composition across the regions. Disentangling these
two causes is a promising avenue for future research.

110

APPENDIX

111

Table 3.1: State-Level Monetary Policy Channels

Variable
Intercept

State Manufacturing Share

State Small Business Share

State Small Bank Share

R-Squared

0.17
(0.31)
0.96
(0.72)
-1.52*
(0.83)
0.00
(0.31)
0.16

The dependent variable is the 2nd quarter cumulative response of state real personal income growth
to a contractionary monetary policy shock. Robust standard errors are in parentheses. * indicates
statistical significance at the 10% level.

112

Table 3.2: State-Level Monetary Policy Response and Industry Shares

Variable
Mining & Logging

Construction

Manufacturing

Trade, Transportation & Utilities

Financial Activities

Services

Government

Intercept

R-Squared

Early Sample (1958-1989) Recent Sample (1990-2008)

6.00
(4.20)
1.32
(5.93)
-2.38**
(1.12)
-0.01
(2.10)
14.91**
(6.53)
-2.31***
(0.77)
-2.73
(1.87)
-0.32
(1.20)
0.41

-8.49
(6.28)
0.72
(8.99)
-5.48
(7.99)
-9.33
(8.93)
-7.50
(10.63)
-7.35
(8.77)
-10.61
(9.86)
7.38
(8.54)
0.34

Robust standard errors in parantheses. ** indicates statistical significance at the 5% level. ***
indicates statistical significance at the 1% level.

113

Table 3.3: State Industry Share Standard Deviations

Industry

M&L Const Manuf TTPU Finance Services Govt

Early Sample

Recent Sample

0.02

0.01

0.01

0.01

0.47

0.04

0.03

0.01

0.01

0.01

0.05

0.04

0.04

0.03

This table shows the standard deviation of employment share in each of the major sectors across
states for the early and recent samples.

114

Figure 3.1: Cumulative Response to Monetary Policy Shock (1958.Q1-2015.Q2)

115

Figure 3.2: Cumulative Response to Monetary Policy Shock (1958.Q1-1992.Q4)

116

Figure 3.3: Cumulative Response to Monetary Policy Shock (1993.Q1-2015.Q2)

117

Figure 3.4: Cumulative Response to Monetary Policy Shock (1984.Q1-2008.Q4)

118

BIBLIOGRAPHY

119

BIBLIOGRAPHY

Adams, R. M., and D. F. Amel (2011): “Market structure and the pass-through of the federal funds

rate,” Journal of Banking & Finance, 35(5), 1087–1096.

Amel, D. F. (1993): “State laws affecting the geographic expansion of commercial banks,” unpub-

lished manuscript, Board of Governors of the Federal Reserve System.

Amidu, M., and S. Wolfe (2013): “The effect of banking market structure on the lending channel:

Evidence from emerging markets,” Review of Financial Economics, 22(4), 146–157.

Araujo, L., and R. Minetti (2011): “On the Essentiality of Banks,” International Economic

Review, 52(3), 679–691.

Ashcraft, A. B. (2006): “New Evidence on the Lending Channel,” Journal of Money, Credit, and

Banking, 38(3), 751–775.

Barakchian, S. M., and C. Crowe (2013): “Monetary policy matters: Evidence from new shocks

data,” Journal of Monetary Economics, 60(8), 950–966.

Beckworth, D. (2010): “One nation under the fed? The asymmetric effects of US monetary policy
and its implications for the United States as an optimal currency area,” Journal of Macroeco-
nomics, 32(3), 732–746.

Berger, A. N., and C. H. Bouwman (2009): “Bank liquidity creation,” The Review of Financial

Studies, 22(9), 3779–3837.

Berger, A. N., A. Demirguc-Kunt, R. Levine, and J. G. Haubrich (2004): “Bank concentration
and competition: An evolution in the making,” Journal of Money, Credit, and Banking, 36(3),
433–451.

Berger, A. N., R. S. Demsetz, and P. E. Strahan (1999): “The consolidation of the financial
services industry: Causes, consequences, and implications for the future,” Journal of Banking
& Finance, 23(2), 135–194.

Berger, A. N., and R. DeYoung (2001): “The effects of geographic expansion on bank efficiency,”

Journal of Financial Services Research, 19(2), 163–184.

Berger, A. N., and G. F. Udell (1995): “Relationship lending and lines of credit in small firm

finance,” Journal of Business, pp. 351–381.

(2002): “Small business credit availability and relationship lending: The importance of

bank organisational structure,” The Economic Journal, 112(477).

Bernanke, B. S., and K. N. Kuttner (2005): “What explains the stock market’s reaction to

Federal Reserve policy?,” The Journal of Finance, 60(3), 1221–1257.

120

Bluedorn, J. C., C. Bowdler, and C. Kochc (2017): “Heterogeneous Bank Lending Responses
to Monetary Policy: New Evidence from a Real-Time Identification,” International Journal of
Central Banking.

Bofondi, M., and G. Gobbi (2006): “Informational barriers to entry into credit markets,” Review

of Finance, 10(1), 39–67.

Boivin, J., M. T. Kiley, and F. S. Mishkin (2010): “How has the monetary transmission mechanism

evolved over time?,” in Handbook of monetary economics, vol. 3, pp. 369–422. Elsevier.

Bonaccorsi di Patti, E., and G. Gobbi (2001): “The changing structure of local credit markets:

are small businesses special?,” Journal of Banking & Finance, 25(12), 2209–2237.

Calomiris, C., and J. Karceski (2000): “Is the bank merger wave of the 1990s efficient? Lessons
from nine case studies,” in Mergers and Productivity, pp. 93–178. University of Chicago Press.
Campbell, J., J. Fisher, A. Justiniano, and L. Melosi (2016): “Forward Guidance and Macroe-
conomic Outcomes Since the Financial Crisis,” in NBER Macroeconomics Annual 2016, Volume
31. University of Chicago Press.

Campbell, J. R., C. L. Evans, J. D. Fisher, and A. Justiniano (2012): “Macroeconomic effects

of federal reserve forward guidance ,” Brookings Papers on Economic Activity, pp. 1–80.

Campbell, J. Y., and J. Ammer (1993): “What moves the stock and bond markets? A variance

decomposition for long-term asset returns,” The Journal of Finance, 48(1), 3–37.

Campbell, J. Y., and R. J. Shiller (1988): “The dividend-price ratio and expectations of future

dividends and discount factors,” Review of financial studies, 1(3), 195–228.

Carlino, G., and R. DeFina (1998): “The differential regional effects of monetary policy,” Review

of Economics and Statistics, 80(4), 572–587.

Cetorelli, N., and L. S. Goldberg (2012): “Banking globalization and monetary transmission,”

The Journal of Finance, 67(5), 1811–1843.

Chen, L., and X. Zhao (2009): “Return decomposition,” Review of Financial Studies, 22(12),

5213–5249.

Chortareas, G., G. Kapetanios, and A. Ventouri (2016): “Credit market freedom and cost

efficiency in US state banking,” Journal of Empirical Finance, 37, 173–185.

Cieslak, A., A. Morse, and A. Vissing-Jorgensen (2016): “Stock returns over the FOMC cycle,”

Available at SSRN 2687614.

Coibion, O., Y. Gorodnichenko, L. Kueng, and J. Silvia (2012): “Innocent bystanders? Mone-
tary policy and inequality in the US,” Discussion paper, National Bureau of Economic Research.

Cole, R. A., L. G. Goldberg, and L. J. White (2004): “Cookie cutter vs. character: The
micro structure of small business lending by large and small banks,” Journal of Financial and
Quantitative Analysis, 39(2), 227–251.

121

Crone, T. M. (2005): “An alternative definition of economic regions in the United States based on

similarities in state business cycles,” Review of Economics and Statistics, 87(4), 617–626.

Crone, T. M., et al. (2007): “A pattern of regional differences in the effects of monetary policy,”

Business Review, (Q3), 9–19.

Davig, T., and J. R. Gerlach (2006): “State-Dependent Stock Market Reactions to Monetary

Policy,” International Journal of Central Banking, 2(4).

Degryse, H., and S. Ongena (2005): “Distance, lending relationships, and competition,” The

Journal of Finance, 60(1), 231–266.

Den Haan, W., S. Sumner, and G. Yamashiro (2002): “Construction of aggregate and re-
gional bank data using the call reports: data manual,” Unpublished manuscript, University of
Amsterdam.

Den Haan, W. J., S. W. Sumner, and G. M. Yamashiro (2007): “Bank loan portfolios and the

monetary transmission mechanism,” Journal of Monetary Economics, 54(3), 904–924.

Ehrmann, M., and M. Fratzscher (2004): “Taking Stock: Monetary Policy Transmission to

Equity Markets,” Journal of Money, Credit, and Banking, 36(4), 719–737.

Eijffinger, S., R. Mahieu, and L. Raes (2017): “Can the Fed Talk the Hind Legs Off the Stock

Market?,” International Journal of Central Banking, 13(1), 53–94.

Evanoff, D. D., and E. Ors (2008): “The competitive dynamics of geographic deregulation in
banking: Implications for productive efficiency,” Journal of Money, Credit and Banking, 40(5),
897–928.

Faust, J., E. T. Swanson, and J. H. Wright (2004a): “Do Federal Reserve policy surprises reveal

superior information about the economy?,” Contributions in Macroeconomics, 4(1).

(2004b): “Identifying VARs based on high frequency futures data,” Journal of Monetary

Economics, 51(6), 1107–1131.

Feroli, M., D. Greenlaw, P. Hooper, F. S. Mishkin, and A. Sufi (2017): “Language after liftoff:
Fed communication away from the zero lower bound,” Research in Economics, 71(3), 452–490.
FungÆ£ovÆ, Z., L. Solanko, and L. Weill (2014): “Does competition influence the bank lending

channel in the euro area?,” Journal of Banking & Finance, 49, 356–366.

Gertler, M., and P. Karadi (2015a): “Monetary Policy Surprises, Credit Costs, and Economic

Activity,” American Economic Journal: Macroeconomics, 7(1), 44–76.

(2015b): “Monetary policy surprises, credit costs, and economic activity,” American

Economic Journal: Macroeconomics, 7(1), 44–76.

Gorodnichenko, Y., and M. Weber (2016): “Are sticky prices costly? Evidence from the stock

market,” The American Economic Review, 106(1), 165–199.

122

G(cid:252)rkaynak, R. S., B. Sack, and E. T. Swanson (2005): “Do Actions Speak Louder Than Words?
The Response of Asset Prices to Monetary Policy Actions and Statements,” International Journal
of Central Banking.

Hanson, S. G., and J. C. Stein (2015): “Monetary policy and long-term real rates,” Journal of

Financial Economics, 115(3), 429–448.

Ippolito, F., A. K. Ozdagli, and A. Perez-Orive (2013): “Is bank debt special for the transmission

of monetary policy? Evidence from the stock market,” Working Paper.

Jansen, D. W., and C.-L. Tsai (2010): “Monetary policy and stock returns: Financing constraints

and asymmetries in bull and bear markets,” Journal of Empirical finance, 17(5), 981–990.

Jayaratne, J., and P. E. Strahan (1998): “Entry restrictions, industry evolution, and dynamic
efficiency: Evidence from commercial banking,” The Journal of Law and Economics, 41(1),
239–274.

Jeon, Y., and S. M. Miller (2003): “Deregulation and structural change in the US commercial

banking industry,” Eastern Economic Journal, 29(3), 391–414.

Kashyap, A. K., and J. C. Stein (1995): “The impact of monetary policy on bank balance sheets,”

in Carnegie-Rochester Conference Series on Public Policy, vol. 42, pp. 151–195. Elsevier.

(2000): “What do a million observations on banks say about the transmission of monetary

policy?,” American Economic Review, pp. 407–428.

Kishan, R. P., and T. P. Opiela (2000): “Bank size, bank capital, and the bank lending channel,”

Journal of Money, Credit and Banking, pp. 121–141.

Kouparitsas, M. (2001): “Is the United States an optimum currency area? An empirical analysis

of regional business cycles,” Federal Reserve Bank of Chicago, Working Paper.

Kroszner, R. S., and P. E. Strahan (1999): “What drives deregulation? Economics and politics
of the relaxation of bank branching restrictions,” The Quarterly Journal of Economics, 114(4),
1437–1467.

Kurov, A. (2012): “What determines the stock market’s reaction to monetary policy statements?,”

Review of Financial Economics, 21(4), 175–187.

Kuttner, K. N. (2001): “Monetary policy surprises and interest rates: Evidence from the Fed

funds futures market,” Journal of monetary economics, 47(3), 523–544.

Laeven, L., and H. Tong (2012): “US monetary shocks and global stock prices,” Journal of

Financial Intermediation, 21(3), 530–547.

Lakdawala, A. (2017): “Decomposing the Effects of Monetary Policy Using an External Instru-

ments SVAR,” Working Paper.

Leroy, A. (2014): “Competition and the bank lending channel in Eurozone,” Journal of Interna-

tional Financial Markets, Institutions and Money, 31, 296–314.

123

Lucca, D. O., and E. Moench (2015): “The Pre-FOMC Announcement Drift,” The Journal of

Finance, 70(1), 329–371.

Maio, P. (2013): “Another look at the stock return response to monetary policy actions,” Review

of Finance, 18(1), 321–371.

Melosi, L. (2016): “Signalling Effects of Monetary Policy,” The Review of Economic Studies,

84(2), 853–884.

Mertens, K., and M. O. Ravn (2013): “The dynamic effects of personal and corporate income

tax changes in the United States,” The American Economic Review, 103(4), 1212–1247.

Minetti, R. (2011): “Informed finance and technological conservatism,” Review of Finance, 15(3),

633–692.

Miranda-Agrippino, S. (2016): “Unsurprising Shocks: Information, Premia, and the Monetary

Transmission,” Unpublished Manuscript, Bank of England.

Nakamura, E., and J. Steinsson (2017): “High frequency identification of monetary non-

neutrality: the information effect,” Quarterly Journal of Economics, forthcoming.

Neuhierl, A., and M. Weber (2017): “Monetary policy slope and the stock market,” Unpublished

manuscript, University of Chicago Booth School of Business.

Olivero, M. P., Y. Li, and B. N. Jeon (2011): “Competition in banking and the lending channel:
Evidence from bank-level data in Asia and Latin America,” Journal of Banking & Finance,
35(3), 560–571.

Owyang, M. T., and H. J. Wall (2003): “Structural Breaks and Regional Disparities in the
Transmission of Monetary Policy,” Federal Reserve Bank of St. Louis Working Paper Series,
(2003-008).

Peek, J., and E. S. Rosengren (1998): “Bank consolidation and small business lending: It’s not

just bank size that matters,” Journal of Banking & Finance, 22(6), 799–819.

Petersen, M. A., and R. G. Rajan (1994): “The benefits of firm-creditor relationships: evidence

from small business data,” The Journal of Finance, 49(1).

Piazzesi, M., and E. T. Swanson (2008): “Futures prices as risk-adjusted forecasts of monetary

policy,” Journal of Monetary Economics, 55(4), 677–691.

Rhoades, S. A. (2000): “Bank mergers and banking structure in the United States, 1980-98,”

Discussion paper, Board of Governors of the Federal Reserve System (US).

Romer, C. D., and D. H. Romer (2004): “A New Measure of Monetary Shocks: Derivation and

Implications,” The American Economic Review, 94(4), 1055.

Romer, D., and C. Romer (2000): “Federal Reserve Information and the Behavior of Interest

Rates,” American Economic Review, 90(3), 429–457.

124

Sapienza, P. (2002): “The effects of banking mergers on loan contracts,” The Journal of Finance,

57(1), 329–367.

Schunk, D. L. (2005): “The differential impacts of monetary policy: Are the differences dimin-

ishing?,” Papers in Regional Science, 84(1), 127–136.

Stiroh, K. J., and P. E. Strahan (2003): “Competitive dynamics of deregulation: Evidence from

US banking,” Journal of Money, Credit, and Banking, 35(5), 801–828.

Stock, J. H., and M. Watson (2012): “Disentangling the Channels of the 2007 to 2008 Recession,”

Brookings Papers on Economic Activity: Spring 2012, p. 81.

Tang, J. (2015): “Uncertainty and the signaling channel of monetary policy,” Discussion paper,

Federal Reserve Bank of Boston.

Weber, M. (2015): “Nominal rigidities and asset pricing,” Working Paper.
Yang, J., and H. Shao (2016): “Impact of bank competition on the bank lending channel of
monetary transmission: Evidence from China,” International Review of Economics & Finance,
43, 468–481.

Zarutskie, R. (2013): “Competition, financial innovation and commercial bank loan portfolios,”

Journal of Financial Intermediation, 22(3), 373–396.

Zou, Y., S. M. Miller, and B. Malamud (2011): “Geographic deregulation and commercial bank
performance in US state banking markets,” The Quarterly Review of Economics and Finance,
51(1), 28–35.

125