Legal Structure, Financial Structure, and the Monetary Policy Transmission Mechanism

Legal Structure, Financial
Structure, and the Monetary Policy
Transmission Mechanism
Stephen G. Cecchetti
ver the past decade, the countries of central
Bank plus the central banks of the eleven monetary union
Europe have become more alike in many
member countries) easier, numerous difficult challenges
ways. As the new members of the European
remain. Primary among these is the making of policy in
OMonetary Union (EMU) prepared for the the face of the possibility that it will have differential
birth of the euro on January 1, 1999, their economic policies
impacts across the countries of the euro area.
became substantially more uniform. All eleven countries in
The task facing the Eurosystem is even more
the new euro area have virtually eliminated inflation and
complex than that facing countries with stable monetary
taken serious steps toward fiscal consolidation.1 As their
regimes, where the measurement of the national and
monetary and fiscal policies have adjusted to meet these
regional impact of policy has already proved to be extremely
common goals, the countries’ business cycle fluctuations
difficult. The creation of the Eurosystem constitutes a
appear to have become more synchronized as well.2 While
regime shift in virtually every sense of the term. The
this makes the job of the Eurosystem (the European Central
introduction of the euro seems sure to prompt adjust-
ments in the economies of the member countries, and
these adjustments will probably alter the relationship
Stephen G. Cecchetti is an executive vice president and the director of research at
between the actions of the central bank and the real econ-
the Federal Reserve Bank of New York.
omy. That is, the monetary transmission mechanism of the
This paper was prepared for the conference “The Monetary Transmission
Process: Recent Developments and Lessons for Europe,” sponsored by the Deutsche
countries in the euro area will change, making the job of
Bundesbank and held in Frankfurt, Germany, on March 26-27, 1999, and is
the new European Central Bank even more difficult than
forthcoming in Deutsche Bundesbank, ed., The Monetary Transmission
it is already. But how quickly will it change, and what
Process: Recent Developments and Lessons for Europe (London:
Macmillan).
will it become?
FRBNY ECONOMIC POLICY REVIEW / JULY 1999
9

---

To answer these questions, we must understand
that the impact of monetary policy innovations varies
the fundamental determinants of the impact of policy
across countries with the strength and scope of the banking
actions on output and inflation. For insight into these
system?
determinants, I turn to the modern views of the monetary
With this in mind, I examine differences in the
transmission mechanism, which assign a central role to
size, concentration, and health of national banking systems,
financial structure. Kashyap and Stein (1997) provide a
as well as in the availability of nonbank sources of finance.
I find, consistent with the most casual observation, that
banking system characteristics vary dramatically across the
countries of the European Union (EU). Furthermore, these
Is there evidence that the impact of monetary
differences do seem to be related to estimated differences in
policy innovations varies across countries with
the impact of interest rate changes on output and inflation.
Countries with many small banks, less healthy banking
the strength and scope of the banking system?
systems, and poorer direct capital access display a greater
sensitivity to policy changes than do countries with big,
healthy banks and deep, well-developed capital markets.
starting point; they focus on the importance of the banking
But this is just the first question. The more
system and go on to emphasize the distributional effects of
important issue facing the Eurosystem is whether the
monetary policy changes. The conventional wisdom has
national banking systems, and the implied sensitivity of
always been that some industries are more sensitive to
each country’s real economy to monetary policy shocks,
interest rate changes than others, and so changes in
will change now that there is monetary union.
policy-controlled interest rates have differential effects
It is easy to assert that European banks will soon
across industries. The view based on financial structure
look like U.S. banks, exhibiting a financial structure and
both formalizes this reasoning and takes it one step further
transmission mechanism similar to the American models.
by noting that some firms are more dependent on banks for
After all, the euro area does resemble the United States, at
financing than others, and that this is true both across and
least superficially. It has a slightly larger population—
within industries. According to this “lending view” of the
292 million for the eleven members of the monetary union
transmission mechanism, monetary policy actions change
relative to 270 million for the United States—and nearly
the reserves available to the banking system, thereby affect-
as high a level of GDP—$6.8 trillion compared with
ing the willingness of banks to lend and, ultimately, the
$8.1 trillion in 1997. The euro area also has a similar
supply of loans. How this mechanism will affect individual
degree of openness to trade, with imports accounting for
firms depends on the financing methods available to them.
slightly more than 10 percent of GDP. These parallels,
Monetary policy has a bigger impact on firms that are
along with the fact that financial technology is easily trans-
reliant on banks for their financing. Furthermore, healthier
ferable across national boundaries, have led a number of
banks will be able to adjust to the policy-induced reserve
observers to conclude that the introduction of the euro may
changes more easily than other banks will.
act as a catalyst, speeding the rate at which financial rela-
The distributional effects implied by the lending
tionships in Europe become like those in the United States.
view of monetary policy transmission have clear implica-
For example, while Dornbusch, Favero, and Giavazzi
tions for the euro area and the Eurosystem. Countries in
(1998, pp. 48-9) do note the possibility for EU-wide
which firms are more bank dependent and banking systems
asymmetries resulting from differences in financial
are less healthy will be more sensitive to the Eurosystem’s
structure, they assert that “the euro will change the way
decisions to change interest rates. This brings me to the
financial markets work, inducing corresponding changes
first question I will address in this paper: Is there evidence
in the monetary mechanism. In addition to pervasive
10
FRBNY ECONOMIC POLICY REVIEW / JULY 1999

---

deregulation already under way and innovation, the intro-
then go on to assert that “legal differences between EU
duction of the euro will revolutionize the financial struc-
states, in particular the lack of some form of ‘European
ture of Europe. Europe will in a short period become more
corporate law,’ also remain important and constitute an
nearly like the USA.” McCauley and White (1997, p. 17)
additional factor of market segmentation” (p. 45). Such
suggest that there may be an acceleration in the rate at
disparities in legal structure can explain important
which securities replace loans on the asset side of bank bal-
economic patterns, and they can be maintained for long
ance sheets and commercial paper replaces deposits on the
periods of time, significantly delaying the harmonization
liability side. They point to a “dramatic potential for assets
of national banking systems.4
to be stripped out of the banking system” and for securities
It is my main contention that the differences in
markets to absorb as much as one-third of the corporate
financial structure across the countries of Europe are a con-
loans now originated in European banks.3 Overall, these
sequence of their dissimilar legal structures. My argument
draws on the work of La Porta, Lopez-de-Silanes, Shleifer,
and Vishny (1997, 1998), who focus on the relationship
between legal structures and finance. They argue that the
It is my main contention that the differences
structure of finance in a country depends on the rights
accorded shareholders and creditors by the laws of that
in financial structure across the countries of
country, as well as on the degree to which these laws are
Europe are a consequence of their dissimilar
enforced. The nature of the laws is, in turn, a product of
the legal tradition on which the civil codes of a country are
legal structures.
based. La Porta et al. establish that the character of a
country’s financial markets depends on the country’s legal
structure. Putting their arguments together with the lend-
commentators are speculating that the increased liquidity
ing view of the monetary transmission mechanism leads to
of European financial markets brought about by monetary
the possibility that it is the legal system in a country that
union will lead to significant consolidation of banks, with
forms the basis for the structure of financial intermediation
mergers at both the national and the international level, as
and, hence, for the impact of monetary policy on output
well as a direct substitution of traded equities and bonds
and prices.
for bank loans.
Table 1 reports the empirical findings that support
Why should we believe that the European finan-
the basic conclusion of the paper. After classifying coun-
cial structure will quickly be transformed into one that
tries by the origin, or “family,” of their legal structure, I
mirrors the one in the U.S. model? Without an explanation
calculate for each family the average level of an index of
for the evolution of these countries’ national financial
monetary policy’s likely effectiveness (based on banking
structures that is based on their existing differences, such
system size, concentration, and health, with a higher value
claims are unconvincing. What accounts for the variation
implying greater effectiveness) and the estimated impact of
in the financial intermediation systems across countries?
an interest rate change on output and inflation (from a
Traditionally, we look to taxes and regulation for an expla-
small-scale structural model). The results suggest that a
nation, and Dornbusch, Favero, and Giavazzi (1998) as
country’s legal structure, financial structure, and monetary
well as White (1998) do mention these. Danthine, Giavazzi,
transmission mechanism are interconnected. The clear
Vives, and von Thadden (1999) identify a number of barri-
pattern is that the predicted effectiveness and its measured
ers to change in national financial structure and note the
impact vary systematically based on the origin of a
importance of the historical path that has brought each
country’s legal system. Countries with better legal pro-
country’s banks to their current state. Danthine et al.
tection for shareholders and debtors (countries with a legal
FRBNY ECONOMIC POLICY REVIEW / JULY 1999
11

---

following section assesses the national banking systems,
Table 1
EFFECTIVENESS OF POLICY AND THE ORIGINS
including measures of overall size, concentration, health,
OF THE LEGAL SYSTEM
and the relative importance of nonbank finance. Overall,
Index of Effectiveness
Impact of Policy
Legal Family
of Monetary Policy
On Output
On Inflation
this analysis allows me to evaluate the likely strength of
English
1.1
-0.45
-0.21
the lending channel across countries. Subsequently, I report
Scandinavian
1.8
-0.52
-0.22
French
2.1
-0.70
-0.20
estimates, for a set of ten countries, of the impact of an
German
2.4
-1.25
-0.49
interest rate increase on output and inflation. These esti-
Notes: The index of effectiveness of monetary policy, from Table 5, is based on
mates follow the pattern that is expected: Countries where
financial structure variables described in the text under the heading “Likely
Strength of the Transmission Mechanism,” with higher values implying a larger
financial structure data suggest that the lending channel
expected impact of interest rate changes on output and prices. The impact of
policy on output and inflation, from Table 6, is a measure of the maximum
should be strong exhibit more sensitivity to monetary
response, in percentage points, to an interest rate movement of 100 basis points,
policy movements. Following the discussion of these
estimated using a small-scale structural model. Countries are classified by the
origin, or family, of their legal structure, and group means are reported based on
findings, I present the data and arguments from La Porta
data for Ireland, the United Kingdom, and the United States (English common
et al. (1997, 1998) on the relationship between legal and
law); Denmark and Sweden (Scandinavian common law); Belgium, France, Italy,
Portugal, and Spain (French civil law); and Germany (German civil law).
financial structures. This allows me to test the prediction
that countries with poor shareholder and creditor protec-
structure based on English common law) have financial
tions and poor law enforcement will have less developed
structures in which the lending channel of monetary trans-
financial systems and greater sensitivity of output and
mission is expected to be less potent; for these countries,
the measured impact of an interest rate change on output
and inflation is lower.
Unless the laws governing shareholder and
The implication is that unless the laws governing
shareholder and creditor rights and the enforcement of those
creditor rights and the enforcement of those laws
laws are harmonized across the members of the European
are harmonized across the members of the
Monetary Union, monetary policy will continue to have a
differential impact. Put slightly differently, it is my belief
European Monetary Union, monetary policy
that the financial structures in the countries of the euro area
will continue to have a differential impact.
will not converge into one large U.S.-style system unless
there are dramatic legislative changes. If such legal har-
monization occurs—that is, if the civil codes protecting
shareholders and creditors are made uniform across the
inflation to interest rate changes. While far from being
countries that have entered the monetary union—then the
definitive, the results are consistent with my main hypoth-
regional variation in the impact of interest rate changes on
esis: Differences in legal systems give rise to variations in
output and inflation should decrease.5 But if legal conver-
national financial structures, and these variations in turn
gence does not occur, financial structure will remain hetero-
lead to divergences in monetary transmission mechanisms.
geneous, and so will the monetary transmission mechanism,
So long as the legal systems of the euro area countries
and the job of the Eurosystem will be to construct appropri-
remain distinct, the impact of interest rate changes
ate policy that takes these asymmetries into account.6
across these countries will differ.
The remainder of this paper provides the building
blocks for this argument. In the next section, I provide a
THEORIES OF THE TRANSMISSION
brief survey of the theories of the monetary transmission
MECHANISM
mechanism, focusing on the importance of financial struc-
A number of comprehensive surveys of the theories of the
ture to an understanding of monetary transmission. The
monetary transmission mechanism have appeared in recent
12
FRBNY ECONOMIC POLICY REVIEW / JULY 1999

---

years. These include Bernanke (1993), Gertler and Gilchrist
The traditional view, which is largely the founda-
(1993), Kashyap and Stein (1994, 1997), Hubbard (1995),
tion for the textbook IS-LM model, is based on the notion
and my own survey, Cecchetti (1995). As a result, I will
that reductions in the quantity of outside money raise real
be brief.
rates of return. This outcome has two effects, the first
All theories of how interest rate changes affect the
directly from interest rates to investment and the second
real economy have a common starting point. A monetary
through exchange rates. An interest rate increase reduces
policy action begins with a change in the level of bank
investment, as there are fewer profitable projects available
reserves. For this to have any real effects at all, there must
at higher required rates of return. A policy action induces a
be nominal rigidities in the economy. Otherwise, a change
movement along a fixed marginal-efficiency-of-investment
in the nominal quantity of outside money cannot have any
schedule. This interest rate channel will be more powerful
impact on the real interest rate. While the ability of the
the less substitutable outside money is for other assets. The
central bank to change the level of bank reserves is not in
exchange rate channel is also familiar from textbook models.
question, the source of the nominal rigidity that allows the
Here, an interest rate increase results in a real appreciation
change in reserves to alter short-run real rates of return has
of the domestic currency, reducing the foreign demand for
been under debate for decades. The current state of this dis-
domestically produced goods. Regardless of whether the
cussion is well summarized by Christiano, Eichenbaum,
transmission mechanism occurs through the interest rate
and Evans (1997). They distinguish three sets of theories:
channel or the exchange rate channel, there is no real need
one set based on sticky wages, a second set based on sticky
to discuss banks. In fact, there is no reason to distinguish
prices, and a third set built on the idea of limited partici-
any of the “other” assets in investors’ portfolios. This is a
pation. The sticky wage and sticky price models, which are
simple two-asset model.
the most familiar, rest on the idea that there are costs to
An important implication of this traditional model
nominal price and wage changes, and so adjustments are
of the transmission mechanism concerns the incidence of
infrequent. In limited participation models, introduced in
the investment decline. Since there are no externalities or
Rotemberg (1984), individuals (households) are unable to
market imperfections, only the least socially productive
adjust their cash balances sufficiently rapidly in response to
projects, those with the lowest rates of return, go un-
changes in the environment—that is, households have a
funded. As a result, the capital stock is marginally lower,
limited ability to participate in financial markets, and so
but, given that a decline is going to occur, the allocation of
must commit themselves to certain portfolio holdings for
the decline across sectors is socially efficient.
relatively long periods of time.7
As most of the surveys cited earlier emphasize, the
The sources of nominal rigidities are relatively
lending view has two parts, one that focuses on the impact
unimportant for the discussion of the mechanism by which
of policy changes on borrower balance sheets and another
interest rate changes have short-run effects on output and
that focuses on bank loans. In both, the effectiveness of
prices, and so I will move directly to a discussion of the
policy depends on capital market imperfections that make
current theories of the transmission mechanism.8 Our
it easier for some firms to obtain financing than others.
current views are based on the work of Bernanke (1983),
Information asymmetries and moral hazard problems,
Bernanke and Blinder (1992), and Bernanke and Gertler
together with bankruptcy laws, mean that the state of a
(1989, 1990). These authors distinguish between the tradi-
firm’s balance sheet has implications for its ability to
tional money view, in which interest rate movements affect
obtain external finance.9 By reducing expected future sales
the level of investment and exchange rates directly, and the
and by increasing the cost of rolling over a given level of
lending view, in which financial intermediaries play a promi-
nominal debt, policy-induced increases in interest rates
nent role in transmitting monetary impulses to output and
(which are both real and nominal) cause a deterioration in
prices. I will describe each of these views in turn.
the firm’s net worth. Furthermore, there is an asymmetry
FRBNY ECONOMIC POLICY REVIEW / JULY 1999
13

---

of information in that borrowers (firms) have better infor-
Models of monetary policy transmission based
mation about the potential profitability of investment
on financial structure suggest a natural place to begin
projects than do creditors (banks). As a result, as the firm’s
looking for sources of cross-country differences in the
net worth declines, the firm becomes less creditworthy
monetary transmission mechanism. The prediction is
because it has an increased incentive to misrepresent the
that overall, the transmission mechanism will be stronger
in those countries where firms are more bank depen-
dent, and where the banking system is less healthy and
less concentrated. In the first instance, firms that have
Models of monetary policy transmission
less direct access to capital markets are unable to blunt
based on financial structure suggest a
the effect of a contraction in bank loans. In the second,
banks themselves have restricted access to nonreservable
natural place to begin looking for sources
deposits and are forced to contract their balance sheets
of cross-country differences in the
by more for a given change in policy. In the next section
of the paper, I examine data on national financial struc-
monetary transmission mechanism.
ture and try to rank countries based on the likely
strength of the transmission mechanism. To the extent
that these cross-country differences are present, then the
riskiness of potential projects—an outcome that will lead
lending view implies that they will persist until the
potential lenders to increase the risk premium they require
financial structures become more uniform.11
when making a loan. The asymmetry of information
makes internal finance of new investment projects cheaper
LIKELY STRENGTH OF THE TRANSMISSION
than external finance.
MECHANISM
More important for the transmission mechanism
In assessing the likely impact of an interest rate change
per se is that some firms are dependent on banks for finance
on output and prices in the various countries of the EMU,
and that monetary policy affects bank loan supply. A
I follow the recent work of Kashyap and Stein (1997) and
reduction in the quantity of reserves forces a reduction in
assemble data on the size and concentration of the bank-
the level of deposits, which must be matched by a fall in
ing systems, along with measures of banking system
loans. Nevertheless, lower levels of bank loans will have an
health, the importance of bank financing, and the size of
impact on the real economy only insofar as there are firms
firms. The indicators are chosen to conform as closely as
without an alternative source of investment funds.
possible to the economic quantities that the lending view
Substantial empirical evidence supports the im-
suggests should be important. The balance sheets of large,
portance of both capital market imperfections and firm
healthy banks are not as sensitive to policy, because
dependence on bank financing. Kashyap and Stein (1997)
reserve contractions can be readily offset with alternative
provide a summary of two types of studies. The first type
forms of finance that do not attract reserve requirements.
suggests that banks rely to a large extent on reservable-
In addition, I examine measures of the development of
deposit financing and that, for this reason, a contraction in
equity and debt markets in the EMU countries. Firms
reserves will prompt banks to contract their balance sheets,
with ready capital market access, which are more likely to
reducing the supply of loans. The second type establishes
be found in countries with extensive secondary securities
that there are a significant number of bank-dependent
markets, will be better insulated from bank loan-supply
firms that are unable to mitigate the shortfall in bank lend-
contractions. Combining these measures, I construct an
ing with other sources of finance. Overall, recent research
index of the probable strength of the monetary transmis-
does imply the existence of a lending channel.10
sion mechanism.12
14
FRBNY ECONOMIC POLICY REVIEW / JULY 1999

---

To assess the importance of small banks in a coun-
the top ten for nearly two-thirds. Overall, Denmark and
try’s financial system, Table 2 reports the number of banks,
Germany have the least concentrated banking systems in
the number of banks per million population, and measures
Europe. By contrast, large banks clearly dominate Sweden,
of concentration for all of the EU countries plus Japan and
Finland, Belgium, the Netherlands, and Greece. The
the United States.13 The data reveal that Austria and
remaining countries are somewhere in between.
Finland have many more banks per capita—126 and 68
What do these findings imply for the strength of
per million people, respectively—than any of the other
the transmission mechanisms in the countries examined?
countries. The remaining countries fall into roughly three
Austria, Germany, and the United States have systems
groups: The United Kingdom, Japan, and the southern
composed of a network of small banks, and so one would
European countries of Spain, Portugal, and Greece have
expect the lending channel to be relatively strong in those
less than 10 banks per million; the United States and
countries. At the other end of the spectrum, Belgium,
Germany have 40 or slightly more; and the remaining
Finland, Ireland, the Netherlands, Portugal, Sweden, and
countries have between 13 and 25.
the United Kingdom all have banking industries domi-
Turning to the concentration measures in the
nated by a small number of relatively large banks, with a
fourth column of the table, it is interesting to note that
modest periphery of small institutions. The remaining
countries with more banks do not necessarily have less
countries—Denmark, France, Greece, Italy, and Japan—
concentrated banking systems. France, for example, with
fall in a middle group.
1,373 banks and just under 60 million people, has a fairly
The weaker a nation’s banking system, the
high concentration ratio: the top five French banks account
stronger the expected impact of policy movements. With
for a sizable 40 percent of total banking system assets and
this in mind, I have collected a set of standard gauges of
banking system health—return on assets, loan loss provi-
sions, net interest margin, and operating costs—and I have
Table 2
calculated a summary rating of overall system soundness
SIZE AND CONCENTRATION OF THE BANKING INDUSTRY,
BY COUNTRY, 1996
(Table 3). Focusing primarily on the return on assets and
Number of Credit
Banks per
Concentration Ratios:
the average Thomson ratings in Table 3 leads to the
Country
Institutions
Million People
Top Five Banks
following rankings: Ireland, the United Kingdom, and
Monetary union members
Austria
1,019
126
48
the United States have the healthiest banks; Austria,
Belgium
140
14
57
Belgium, Germany, the Netherlands, Spain, Denmark,
Finland
350
68
78
France
1,373
24
40
and Greece are second; Finland, France, Italy, Portugal,
Germany
3,517
43
17
and Sweden are third; and Japan is alone at the bottom.
Ireland
62
18
41
Italy
937
16
25
Finally, I turn to the availability of nonbank
Netherlands
172
11
79
Portugal
51
5
76
finance for firms in EU and other countries. The relevant
Spain
313
8
44
data are reported in Table 4. Following Kashyap and Stein
Members of the EU
(1997) and La Porta, Lopez-de-Silanes, Shleifer, and Vishny
not in EMU
Denmark
117
22
17
(1997), I examine the number of publicly listed firms, the
Greece
20
2
71
extent of secondary equity and debt markets, and the ratio
Sweden
124
14
90
United Kingdom
478
8
28
of bank loans to all forms of finance. Although these are
crude measures of access to external finance, they are infor-
Other countries
Japan
556
4
30
mative. As in the case of Table 2, the countries can be
United States
10,803
40
17
divided into three groups. Austria, Ireland, Italy, Portugal,
Sources: See the Data Appendix.
and Greece appear to have the least well developed external
Note: Concentration ratios are calculated as the percentage of each country’s
bank assets accounted for by the five largest banks.
capital markets. They have small equity and bond markets,
FRBNY ECONOMIC POLICY REVIEW / JULY 1999
15

---

Table 3
MEASURES OF BANKING INDUSTRY HEALTH, BY COUNTRY, 1996
Percent
Return on Assets
Loan Loss Provisions
Net Interest Margin
Operating Costs
Average Thomson Rating
Country
(1)
(2)
(3)
(4)
(5)
Monetary union members
Austria
0.38
0.59
1.67
2.45
2.38 (4)
Belgium
0.52
0.17
1.41
1.67
2.00 (6)
Finland
0.50
0.78
2.07
3.05
2.83 (3)
France
0.36
0.24
1.43
1.84
2.28 (16)
Germany
0.44
0.18
1.24
2.19
1.97 (19)
Ireland
1.57
0.17
3.36
3.32
1.83 (3)
Italy
0.33
0.62
2.32
3.19
2.57 (15)
Netherlands
0.75
0.26
2.06
2.48
2.10 (5)
Portugal
0.91
0.42
2.60
3.80
2.30 (5)
Spain
0.76
0.32
2.20
2.69
1.79 (11)
Members of the EU not in EMU
Denmark
0.91
0.11
1.28
0.97
2.33 (3)
Greece
1.11
0.18
1.98
2.77
2.50 (6)
Sweden
1.28
0.25
1.90
1.77
2.50 (5)
United Kingdom
1.28
0.18
2.15
2.42
2.04 (23)
Other countries
Japan
0.01
0.75
1.17
1.03
3.32 (44)
United States
1.42
0.10
2.68
3.51
1.73 (344)
Sources: See the Data Appendix.
Notes: Except for the Thomson ratings, all figures in the table are calculated as a percentage of total bank assets. In column 5, the number of banks rated by Thomson in
each country and used to compute the average appears in parentheses.
Table 4
IMPORTANCE OF EXTERNAL AND BANK FINANCE BY COUNTRY, 1996
Number of Publicly
Publicly Traded Firms
Market Capitalization as a
Corporate Debt as a
Bank Loans as a Percentage
Traded Firms
per Capita
Percentage of GDP
Percentage of GDP
of All Forms of Finance
Country
(1)
(2)
(3)
(4)
(5)
Monetary union members
Austria
106
13.15
15
46
65
Belgium
139
13.68
45
60
49
Finland
71
13.87
50
34
39
France
686
11.75
38
49
49
Germany
681
8.32
29
58
55
Ireland
76
21.59
18
13
80
Italy
217
3.78
21
37
50
Netherlands
217
13.97
96
48
53
Portugal
158
16.11
23
19
62
Spain
357
9.09
42
11
58
Members of the EU not in EMU
Denmark
237
45.06
41
105
25
Greece
245
23.44
20
3
48
Sweden
229
25.90
99
73
32
United Kingdom
2,433
41.39
150
45
37
Other countries
Japan
2,334
18.56
67
39
59
United States
8,479
31.94
111
64
21
Sources: See the Data Appendix.
Notes: Market capitalization is the year-end value of firms listed on major exchanges. For the United States, three exchanges are used; for Japan, eight; and for each of the
remaining countries, one.
16
FRBNY ECONOMIC POLICY REVIEW / JULY 1999

---

and bank loans account for a high percentage of firm
in which small banks are relatively important, the banking
financing. By contrast, Belgium, Denmark, Sweden, the
systems are less healthy, and firms have little access to non-
United Kingdom, and the United States all have substantial
bank sources of finance. The opposite is true of Belgium,
secondary capital markets, and banks are a less important
Ireland, and the Netherlands, where the banking systems
source of finance. The remaining six countries are some-
are large and healthy and nonbank finance is readily avail-
where in between these two groups.
able; in these countries, interest rate movements would be
Table 5 summarizes the material in Tables 2-4 and
expected to have a more muted impact.14
suggests the overall relative strength of monetary policy in
The conclusions of this section could be criticized
the fourteen EU countries, Japan, and the United States.
as applying only to the pre-EMU period. But will the
The final column, “Predicted Effectiveness of Monetary
introduction of the euro be a catalyst for the harmonization
Policy,” reports a measure of the effects of monetary policy
of financial structure across the EMU? I take this question
on output and inflation, where higher values suggest a
up in more detail later, but at this point I will simply
stronger lending channel and therefore a larger impact.
mention that the recent European Central Bank (1999)
Overall, the pattern is very similar to the one reported in
report Possible Effects of EMU on the EU Banking Systems in
Kashyap and Stein (1997, Table 6). Most important, the
the Medium to Long Term provides very little evidence to
predicted effects of interest rate movements vary greatly
suggest that an increase in either international banking
across countries. For example, looking at the EMU coun-
competition or securitization and disintermediation will
tries, one would expect that a given interest rate change
occur quickly.
would have the most impact in Austria and Italy, countries
MEASURING THE IMPACT OF POLICY
ON OUTPUT AND PRICES
Testing the proposition that the banking system’s concen-
Table 5
SUMMARY OF FACTORS AFFECTING THE STRENGTH
tration, health, and importance have a material impact on
OF THE MONETARY TRANSMISSION MECHANISM
the monetary transmission mechanism requires an estimate
Importance
Availability of
Predicted
of the effects of an interest rate change on output and
of Small
Bank
Alternative
Effectiveness of
Banks
Health
Finance
Monetary Policy
prices. Numerous studies report such estimates for some or
Country
(1)
(2)
(3)
(4)
Monetary union members
all of the countries of the EU. These include Gerlach and
Austria
3
2
3
2.67
Smets (1995), who estimate a three-variable structural vec-
Belgium
1
2
1
1.33
Finland
1
3
2
2.00
tor autoregression based on long-run restrictions; de Bondt
France
2
3
2
2.33
Germany
3
2
2
2.33
(1997), who presents estimates of the impact of policy on
Ireland
1
1
3
1.67
output and prices for Germany, France, Italy, the United
Italy
2
3
3
2.67
Netherlands
1
2
2
1.67
Kingdom, Belgium, and the Netherlands that are based on
Portugal
1
3
3
2.33
the work of other authors; Dornbusch, Favero, and
Spain
2
2
2
2.00
Members of the EU
Giavazzi (1998), who report estimates of the impact of
not in EMU
policy on output and prices derived from both small
Denmark
2
2
1
1.67
Greece
2
2
3
2.33
vector-autoregressive models and large structural models,
Sweden
1
3
1
1.67
United Kingdom
1
1
1
1.00
for Italy, Germany, France, Spain, Sweden, and the United
Other countries
Kingdom; Kieler and Saarenheimo (1998), who study
Japan
2
4
2
2.67
France, Germany, and the United Kingdom, concluding
United States
3
1
1
1.67
that the transmission mechanism is not significantly differ-
Notes: Column 1 is based on Table 2; column 2, on Table 3; and column 3, on
Table 4. Column 4 is an average of columns 1, 2, and 3.
ent across the three countries; and Vlaar and Schuberth
FRBNY ECONOMIC POLICY REVIEW / JULY 1999
17

---

(1998), who examine money demand functions for fourteen
vary dramatically across countries. Looking at the impact
EU countries; Ehrmann (1998), who estimates structural
of interest rate movements on output, note that for France
vector autoregressions for thirteen countries and finds con-
and Germany, the peak impact is nearly twice what it is in
siderable differences in the intensity of the response of
the remaining European countries, and fifteen times the
estimated impact in the United States. The impact of
policy on inflation also varies substantially.
Table 6 reports the maximum impact of a 100-
Testing the proposition that the banking system’s
basis-point monetary contraction on output and inflation
concentration, health, and importance have a
for all of the countries for which I have estimates. I also
include a measure of the timing of the impact—the quarter
material impact on the monetary transmission
at which the maximum effect occurs. The final column in
mechanism requires an estimate of the effects of
the table presents a measure of the ratio of the average
output response to the average inflation response. This
an interest rate change on output and prices.
measure is related to the sacrifice ratio because it is roughly
the output loss for an inflation decline of 1 percentage
point over a horizon of approximately three years. Unfortu-
output and prices to monetary shocks across countries;
nately, these estimates are not terribly precise, a point that
and Cecchetti and Rich (1999), who look at a simple
is clear from the results in Ehrmann’s paper,18 and so we
two-variable system for Australia, Canada, France, Italy,
should not take some of the numbers too seriously.
Switzerland, the United Kingdom, and the United States,
and find large differences in the implied impacts.
SYSTEMATIC DIFFERENCES IN NATIONAL
Each of these studies has advantages and disadvan-
LEGAL SYSTEMS
tages. Overall, I have chosen to examine the results
If differences in financial systems are creating the cross-
reported by Ehrmann (1998). The appeal of Ehrmann’s
sectional variation in the transmission mechanism, it is
approach is that it yields a series of estimates, all based on
natural to look for the causes of these differences. As noted
the same methodology, for nearly the full set of EU coun-
earlier, La Porta, Lopez-de-Silanes, Shleifer, and Vishny
tries. Ehrmann uses techniques devised by King, Plosser,
(1997, 1998) have examined the relationship between a
Stock, and Watson (1991). In effect, he identifies monetary
country’s legal system and its financial system. The
shocks using a combination of long-run and short-run
premise of their work is that investors provide capital to
restrictions. The methods are described both in his paper
firms only if the investors have the ability to get their
and in Cecchetti, McConnell, and Perez Quiros (1999). For
money back. For equity holders, this means that they must
each country, the model has either four or five variables,
be able to vote out directors and managers who do not pay
including output, inflation, and an interest rate, and—
them. For creditors, this means that they must have the
with the exception of Germany—an exchange rate. When a
authority to repossess collateral. In addition to having
fifth variable is present, it is either a second interest rate or
nominal legal rights, these groups must also have confi-
a commodity price index.15
dence that the laws will be enforced.
The chart plots the responses of output and infla-
La Porta et al. (1997, 1998) collect data on the
tion to an interest rate movement of 100 basis points for
legal systems in forty-nine countries. They show that all
ten EU countries and the United States.16 These ten coun-
of these legal systems belong to one of four families:
tries are the ones for which Ehrmann is able to generate
English common law, French civil law, Scandinavian civil
consistent and plausible results.17 As is clear from these
law, and German civil law. With regard to shareholder
plots, the point estimates of the impulse response functions
rights—specifically, the ability of shareholders to vote
18
FRBNY ECONOMIC POLICY REVIEW / JULY 1999

---