This article was downloaded by: [Fudan University]

On: 15 May 2015, At: 10:40

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Regional Studies
Publication details, including instructions for authors and subscription
information:
http://www.tandfonline.com/loi/cres20

Regional Finance: A Survey

a b
Sheila C. Dow & Carlos J. Rodrguez-Fuentes
a
Department of Economics , University of Stirling , Stirling, FK9 4LA, UK
b
Department of Applied Economics , University of La Laguna , S/C
Tenerife, 38071, Spain
Published online: 18 Aug 2010.

To cite this article: Sheila C. Dow & Carlos J. Rodrguez-Fuentes (1997) Regional Finance: A Survey, Regional
Studies, 31:9, 903-920, DOI: 10.1080/00343409750130029

To link to this article: http://dx.doi.org/10.1080/00343409750130029

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content)
contained in the publications on our platform. However, Taylor & Francis, our agents, and our
licensors make no representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views expressed in this
publication are the opinions and views of the authors, and are not the views of or endorsed
by Taylor & Francis. The accuracy of the Content should not be relied upon and should be
independently verified with primary sources of information. Taylor and Francis shall not be liable
for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other
liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in
relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial
or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use
can be found at http://www.tandfonline.com/page/terms-and-conditions
 Regional Studies, Vol. 31.9, pp. 903 920

Regional Finance: A Survey

S H E I L A C. D OW * and C AR L O S J. RO D R I G U E Z -F U E N T ES
*Department of Economics, University of Stirling, Stirling FK9 4LA, UK
Department of Applied Economics, University of La Laguna, S/C Tenerife 38071, Spain

I N T RO D UC T I O N macro literature seems to be a regional extension of

the national discussion about whether or not money
The role of money and nancial markets in the regional
matters.1
economy has been attracting increasing attention in
We start with a review of the treatment of money
recent years. It has been a topic of interest for longer
in regional economics, and then turn to the much
in the US literature, re ecting the regional concerns
larger literature in which macro-monetary theory is
of a federal state. But now European nancial integra-
Downloaded by [Fudan University] at 10:40 15 May 2015

applied to regional economies.

tion and more general political changes have raised the
importance of the topic for Europe. The relevant
literature is now suYciently large and diverse that a
T H E M O N E TA RY D I M E N S I O N I N
survey is warranted.
T RA D I T I O N A L RE G I O N A L A N A LY S I S
While some contributions to the regional nance
literature have arisen from regional economics, the R I C H A R D S O N , 1972, 1973, has oVered some inter-
majority have arisen from macro-monetary economics. esting clues as to why regional economics has usually
This lack of interest in nancial variables within neglected any role for money and nancial variables.
regional economics has often led regional economists First, regional economists have borrowed too freely
to belittle the power of money in explaining regional from growth theory (R I C H A R D S O N , 1973, p. 14).
income diVerences. Regional economists often assume Theories which assume free and costless movement of
money neutrality, so that the entire analysis may be labour and capital among regions, full and costless
conducted in real terms. In other cases regional econ- availability of information, etc. thereby assume away
omists have relied upon the assumption of perfect any regional role for money.
interregional capital mobility. The regional money Furthermore, even when money has been given
supply is then endogenous, but cannot aVect any real some role to play in the determination of national
variable since interregional monetary ows simply mir- income and employment, it has been suggested that
ror regional diVerences in real terms. both the open character of regional economies and the
Therefore monetary variables, if included at all, have absence of regional monetary tools (ibid., pp. 9 13),
usually been considered exogenously determined at the leave regional economies without any monetary iden-
national level, with the regional weighting mirroring tity. Therefore, if money is non-neutral, it is so only at
regional real economic diVerences. Therefore, money the national level. Regional monetary analysis therefore
and monetary ows have been considered as the result corresponded with global monetarism since regions,
of regional economic diVerences rather than a factor like all small open economies,2 face a horizontal supply
which might have played a role in their widening. of money at some interest rate which, in turn, is xed
Just as regional economists have underestimated the in national, or international, markets. Providing that
power of money and nancial factors in the explanation nancial capital ows freely from one region to another,
or widening of regional-income diVerences, so also interregional monetary ows mirror real ows unless
macro-monetary economists have commonly over- some market failure inhibits such an accommodative
simpli ed, if not misunderstood altogether, the real behaviour. All this explains why traditional models of
signi cance of the regional composition of their regional income determination, such as neo-classical
models. While there is a large empirical monetary models, cumulative causation models, I-O models and
literature with a regional dimension, what is meant by multisector models, have excluded monetary variables.
regional application is a regional data-set. There exists Of these traditional models, only export base and
considerable parallelism between those works con- econometric models have included some kind of mon-
cerned with the eVect of monetary policy on the etary variable in their speci cation. In the export base
national economy and those concerned with the models, as regional exports explain regional income
regional impact of such policy. Therefore, the regional growth and as the former may be aVected by national

0034-3404/97/090903-18 1997 Regional Studies Association

 904 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
Table 1. Literature on regional impact of monetary policy theories in a regional setting. These contributions have
Reduced form B E A RE , 1976; C O H EN and M AE SHI RO,
been grouped into six blocks (see Tables 1 and 2). In
models 1977; T OAL , 1977; M AT HUR and the following subsections we provide a more detailed
S T EIN, 1980, 1982, 1983;G AR RI SON account of each group.
and K O RT, 1983; K O ZLOWSKI , 1991 The rst three blocks of this literature (see Table
Large regional macro F I S H K I N D, 1977; M ILL ER , 1978; 1) explicitly address the regional impact of national
models G A RR IS ON and C H AN G, 1979; and
C HAS E E C O N O M ETR IC A SS O C I ATES , monetary policy. The focus here has mostly been on
1981 how national monetary changes aVect the real regional
DiVusion of open S C O TT, 1955; L AWR EN C E , 1963; economy, with the predominant approaches being
market operations B RYAN, 1967; R UF FI N, 1968; B ARTH those within the neo-classical synthesis: the monetarist
et al., 1975; T H UR S TO N, 1976; and
M C P HETER S , 1976
and neo-classical Keynesian.
As at the national level, monetarists have also tried
to get some regional evidence, by means of reduced
form models, that money and monetary policy directly
income growth, regional income could be indirectly
aVected by national monetary policy via its eVect cause business cycles. Neo-classical Keynesians have
on national income growth. Regional econometric instead used large regional structural models to show
models, in turn, have usually included some monetary that monetary policy aVects the economy in a more
variables in their speci cation, either in simple equa- indirect way, at the regional level as well as at the
tions where regional income is linked to some national national level. The third group of papers focuses on
Downloaded by [Fudan University] at 10:40 15 May 2015

variables (such as GDP, interest rates or money supply), the regional eVects of national open market operations.
or in simultaneous models, where interest rates are A large literature also considers regional nancial
very often included as an exogenous variable (see, for variables and markets themselves. Table 2 oVers a
example C Z A M A N S K I , 1969; and G L I C K M A N , 1977, classi cation of this broader literature into three diVer-
1980a, 1980b). ent categories. While many papers extend over more
As R I C H A R D S O N , 1973, pp. 12 13, has pointed than one classi cation, they have been assigned to a
out then, traditional regional economics has not been category according to their major contribution.
able to make any signi cant contribution to discussions First there is the literature focused on how monetary
of national monetary policy. However, the macroecon- multipliers are modi ed when interregional nancial
omic literature has evolved to consider the implications and trade ows are included. The second group is
of the regional composition of the national economy. made up of the literature which has tried to estimate,
And it is in this context that most of the regional rather than explain, interregional nancial ows. The
monetary literature has emerged. We consider this richest literature is in the third group, dealing with
literature in the next section. regional nancial markets. Two sub-categories are dis-
tinguished. The first deals with the issue of regional
interest rate diVerentials whereas the second looks at
T H E RE G I O N A L D I M E NS I O N I N the factors determining regional availability of credit.
M A C RO - M O NE TA RY M O D E L S
Although all these papers have a common subject
In this section we attempt rst to group a large number matter, they clearly diVer in theoretical approach, rang-
of papers which test some competing macro monetary ing from neo-classical explanations of regional diVer-

Table 2. Regional nancial literature

Regional monetary multipliers D EI S S , 1978; D OW, 1982; M O O RE e t a l ., 1 9 85 ; an d G O O D H ART, 1989
Interregional nancial ow of G I LBERT, 1937 1938; H ARTL AN D, 1949; B OWS H ER et al., 1957; L I EBER SON and S CH WI R I AN,
funds 1962; L EES , 1969; A LVAREZ- L L AN O and A N D R EU, 1978; F ER N A N D EZ and A N DR EU, 1978;
C A ST EL L S and S I C ART, 1980; B AN C O D E B ILBAO, 1980; S H O RT and N IC H OL A S , 1981; and
C A R L IN O and L AN G, 1989
Regional nancial markets:
a. Interest rate diVerentials H EN DER S O N, 1944; C ARR , 1960; E DWARDS, 1964; D A V IS and B A N KS, 1965; S C H A AF, 1966;
M E YER , 1967; W I N GE R , 1969; S T R AS ZH E IM , 1971; P ET ERS O N, 1973; C EBU LA and Z A H A RO F F,
1974; J AM ES , 1976; O S TAS , 1977; R O C KOFF, 1977; H E NDER S H OT T and K I DW EL L , 1978;
A SP I N WA L L , 1979; K EL E HER , 1979; H U TCH IN S ON and M C K ILL O P, 1990; M C K I L L O P and
H UT CHIN SON, 1990; DA M IC O et al., 1990; and F AINI et al., 1993
b. Regional credit availability D REE SE , 1974; K EL EHER , 1977; R O B ERTS and F I SH KI N D, 1979; M O O RE and H ILL , 1982;
A L L EN and P R IC E , 1984; K AN N A N, 1987; D OW, 1987a, 1987b, 1987c, 1988, 1990, 1992, and
1996; H A RR IGAN and M C G RE G O R , 1987; C H ICK and D OW, 1988; M O O R E and N AGU R NEY,
1989; S A MO LYK , 1989, 1991 and 1994; H U TC H I NSO N and M C K IL L O P, 1990; M C K ILL O P and
H UT CHIN SON, 1990; H U G H ES , 1991, 1992; B I A S , 1992; A M OS , 1992; A M O S and W IN G EN D ER ,
1993; R O DR I G U EZ- F UE NT ES , 1993; G R EE NWALD et al., 1993; F AINN I et al., 1993; M ES S OR I ,
1993; P O RTEO US , 1995; and C H IC K , 1993
 Regional Finance: A Survey 905
entials in interest rates to the new Keynesian literature consists of state currency holdings, plus state demand
on credit rationing, or the post Keynesian literature on deposits. The deposit totals are estimated by applying
money and credit. We now discuss each of these the national currency-demand deposit ratio to national
literatures in turn. demand deposits (ibid., p. 674). This is a common
method for dealing with the absence of regional nan-
cial data, but it entails theoretical assumptions. Cur-
The regional impact of monetary policy: some reduced-form
rency is held for transactions purposes, but deposit
models
demand may re ect other motives, such as the precau-
While S C OT T, 1955, and L AWR E NC E , 1963, raised tionary and speculative motives. The estimation
the issue of assessing empirically the regional impact of method therefore rules out any regional diVerence in
monetary policy, B E A R E s, 1976, contribution was the liquidity preference, i.e. it is assumed that deposits are
rst to receive signi cant attention. His paper was also only held for transactions purposes, and thus bear
regarded as a further step in the development of the a close relationship with regional output. Much of
monetarist theory of business cycles, this time at the what is being tested for is therefore already assumed in
regional level. The theoretical roots of Beares work the data estimation.
are to be found in F R I ED M A N and M E I S EL M A N , The paper by K O Z L OWS KI , 1991, takes a diVerent
1963, and A N DE R S EN and J OR DAN , 1968. These two approach by attempting to show that the forecasting
papers not only set up the basis of the so-called St Louis performance of regional models of leading indicators
equation, but also provided a theoretical framework for is improved by incorporating a national monetary vari-
most of the monetarist argument on the impact of able. He compared the performance of four leading
Downloaded by [Fudan University] at 10:40 15 May 2015

monetary policy, either at national or regional level. indicator models for Detroit, South Carolina, Toledo
Beare chose to test the monetarist argument that and Wisconsin, with and without a national monetary
business cycles are due to monetary shocks at the indicator. The results showed that, when money was
regional level because: `if money contributes at least to included, forecasting performance was improved. The
some extent to uctuations in national activity levels, implication is that the national money supply should
then it must also contribute to uctuations in the be included in regional models because it is one of the
activity levels of the diVerent regions of a national causes of regional business cycles.
economy (B EA R E , 1976, p. 57). Beare estimated a Within the monetarist Keynesianism debate over
monetarist reduced-form model for the Prairie Prov- the explanation of regional business cycles, some con-
inces of Canada (Manitoba, Saskatchewan and Alberta) tributions, such as M AT H U R and S T E I N, 1980, 1982,
with annual data for the period 1956 71. Beares 1983, pointed out the dangers and limitations of fol-
model, speci ed both in nominal and real terms, was lowing simple reduced-form approaches such as the St
a simple extension of the St Louis equation to a regional Louis equation. M ATH UR and S T EI N , 1980, were
setting. Beare concludes that his model con rms the mainly concerned with the bias problem which arises
importance of national money in the determination when using such reduced-form models. They support
of regional income, with diVerential regional impact their hypothesis both theoretically and empirically.
explained by regional diVerences in income or wealth Empirically they tested a reduced-form model similar
elasticity of demand for regional output. Beare claims to that of B EA R E , 1976. The model was estimated for
that his results support the monetarist view: `[T]he eight US regions, for two sample periods, 1952:I to
initial eVects of a monetary change are principally on 1968:II and 1952:I to 1976:IV. The results showed
real output rates, . . . the long-run eVects are principally money multipliers as highly signi cant compared to
(and perhaps totally) felt on the price level (B E A R E , scal multipliers, consistent with the monetarist
1976, p. 58). But real money holdings seem to have a account. However, it was emphasized that both multi-
strong eVect on real regional income in almost every pliers showed a high degree of instability, raising scepti-
estimation over the full period. That is, money seems cism about the use of reduced-form models at the
to aVect regional real income over a 15-year time-span. regional level.
This in fact challenges the monetarist view that money G A R R IS ON and K O RT, 1983, was a response both
is neutral in the long run. to Beares monetarist explanation of the regional busi-
C O H EN and M A ES H I RO s, 1977, paper was also ness cycle and also to the scepticism showed by Mathur
designed to test the monetarist view of business cycles and Stein regarding the use of reduced-form models.
at the regional level. They did so by estimating two In order to demonstrate both the power of scal
equations for the US regions for the period 1948 71, variables in the explanation of regional business cycles
one with regional income the dependent variable and and the usefulness of reduced-form models, they esti-
regional money independent, and the reverse in the mated the in uence of national scal and monetary
second equation. The authors concluded that their variables on regional employment by states in the US
model supported the monetarist view, i.e., that income for the period 1960:I to 1978:IV. The results con rmed
is the dependent variable. But data issues raise some that both monetary and scal variables in uence real
doubts. For example, the regional money variable activity. They also argued that the instability found by
 906 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
M AT H U R and S T EI N , 1980, 1982, was due to the Some large regional macro models analysing the regional
wrong selection both of independent variable and impact of monetary policy
period. However, these criticisms were rejected by
Whereas monetarists have mainly addressed the issue
M AT H U R and S T EI N , 1983.
of the regional impact of monetary policy by means of
In most of the approximations reviewed so far there
some kind of reduced-form model, Keynesians have
remain some black boxes regarding the explanation of
instead chosen to develop some kind of large regional
the relationship between money and output. The gen-
macro model. One group of models has been aimed at
eral standpoint in all these monetarist models has been
assessing the regional side eVects of national nancial
that, as economic growth depends on real factors, any
variables mainly interest rates (although they have
monetary change (which is assumed to be exogenous
not directly addressed the matter of the regional impact
to the system) is only able to disturb economic activity,
of monetary policy).4 A second group of papers have
creating instability in the short-run,3 or in ation in the
directly addressed the issue of the regional impact of
long-run. Others, on the contrary, have maintained
monetary policy, either from a monetarist or Keynesian
that scal variables are also important in the explanation
perspective.
of business cycle.
Large regional macro models built up on Keynesian
No solution to this debate has arisen from the
assumptions have mainly recognized two diVerent
empirical evidence. B E A R E s, 1976, and G A R R I S O N
and K O RT s, 1983, papers, for example, show how channels through which national monetary policies
examination of the evidence has left the debate open have aVected regional economies. The rst channel has
been the eVect of money on national income, which
Downloaded by [Fudan University] at 10:40 15 May 2015

because it has given support to both explanations.

is a major determinant of regional income. The second
However, further empirical studies would not yield a
de nitive conclusion for a variety of reasons, some of channel has been the eVect of national interest rates on
which are listed below: regional expenditure. Monetarist models have taken a
diVerent approach, being more concerned with how
? as M ATH UR and S T EI N , 1980, 1982, 1983, have national monetary policy has aVected the regional
pointed out, reduced form models may not be distribution of money and, therefore, how monetary
capable of discriminating between competing sides policy may aVect regional business cycles. In what
of the debate follows, some brief comments on both kind of models
? diVerences in data-sets and econometric tools add will be oVered.
noise to the debate Within the Keynesian tradition, F I S H KI N D, 1977,
? there may not exist a monetary theory of general developed a short-run export-base model for the
applicability; the suitability of any theory may well Indiana economy for the period 1958 1973. The
depend on institutional factors, which diVer from model is not fully presented in the article, although the
case to case. author states that it is composed of some 34 equations,
of which 17 are stochastic. According to Fishkind, the
But a more fundamental issue arises if money is not model contains all the three channels of monetary
exogenous to the (national) economy as both monetar- policy: (1) the cost of capital, which is included in the
ists and neo-classical Keynesians suggest, but is the housing investment equation: (2) the availability of
result of a complex process of interaction between capital, also included in the housing investment func-
monetary authorities interventions and private agents tion; and (3) the wealth eVect.
responses to those interventions. If that were the case, In order to test the diVerential eVect of US national
how should the correlation between money and monetary policy on the Indiana economy, Fishkind
income be understood? In fact, money and income compares the behaviour of some regional economic
would then show a strong correlation over the cycle, indicators with the national equivalents during periods
because credit is what nances production and money of tight monetary policy (1969 70) and easy monetary
is credit-driven. From this theoretical point of view, policy (1971 72).5 The results showed that, in the
money rather than being the cause, would be the eVect former, the Indiana economy experienced slower
of the cycle. From this perspective, the correlation growth than the US. However, during easy money
between money and income, either at national or times, Indiana grew at the same rate as the US. This
regional level, would have a diVerent interpretation to asymmetrical behaviour was explained by the relative
the one given by either monetarist or neo-classical composition of the state economy.
Keynesian models. G A R R IS ON and C H A NG , 1979, estimated a regional
Furthermore, even if money were the single cause Keynesian model built on export-base theory for eight
of economic instability, it would remain diYcult to put US regions with quarterly data for the period 1969:I
into practice the monetarist proposal of controlling to 1976:I. Garrison and Chang concluded, in line
money supply in order to avoid business cycles, if with Keynesian theory, that both monetary and scal
money is no longer under the control of the monetary variables in uence economic activity.
authorities. They also worked out the scal and monetary elasti-
 Regional Finance: A Survey 907
cities of manufacturing income and found that regions Cleveland. Millers explanation of the workings of
with higher concentrations in durable goods manufac- the model is as follows. The process begins with an
turing experienced larger responses to changes in all exogenous monetary policy manipulation by the cen-
variables. Therefore, they concluded, monetary and tral monetary authority: the Fed. This monetary
scal policy are likely to have diVerent regional eVects manipulation changes the ow of net source to a region
due to the regional diVerences in economic structure. which is, in turn, ampli ed by the multiplier process.
Regions having a higher concentration of durable This change in the regions money supply generates
goods manufacturing will be more aVected than those changes in the regions level of economic activity. This,
where agriculture and mining are more important. in turn, in uences interregional economic relationships
C H A S E E C O NO M ET R I C A S S O C I AT E S , 1981, (interregional trade in goods and services) which then
arrived at the same results as G A R R IS ON and C H A N G , generate new interregional monetary ows. These
1979, by means of the estimation of a regional model monetary ows change the regions net source base
for some eight US regions, four rural and four urban. and the process starts again. Miller recognized that: `the
The model is made up of 164 equations, 114 identities explanatory power of these reduced form equations
and 38 exogenous variables. They concluded that urban was inadequate, . . . and [this low signi cance] suggested
regions seem to be more aVected by tight monetary the need for additional explanatory variables, such as
policies than rural ones, due to diVerences in elasticities scal explanatory variables (ibid., p. 141). But, a page
between diVerent economic sectors. Two exercises were later, he concludes `the regional pattern of monetary
carried out. First, the eVect of a tight monetary policy policy . . . supported a rate of growth in nominal
was simulated. Second, the eVect of a redistribution of personal income in the rest of the country which was
Downloaded by [Fudan University] at 10:40 15 May 2015

credit from urban to rural regions was estimated. Both greater than the rate of growth in nominal personal
exercises gave support to the results already noted, i.e. income in the Northeast region (ibid., p. 142).
tight monetary policy had a greater eVect on urban So far we have reviewed some works which have
regions. tried to assess the issue of the regional impact of
M I L L E R s, 1978, book oVers a monetarist analysis monetary policy empirically, using either some kind of
of the regional impact of monetary policy. He sets reduced-form model, or some kind of large regional
out a short-run, two-region, macroeconomic static macroeconomic model. Some of them have given
multiplier model which, combining the global monet- empirical support to the monetarist view of the
arist approach to the balance of payments and a regional (regional) business cycle. Others have proved that scal
speci cation of the money supply, allowed him to test variables also matter.
the channels through which monetary policy aVects But despite the diVerences, there are also some
regional economies. His model thus goes further than important elements held in common. First, the money
the global monetarism literature by including a regional supply is considered to be perfectly exogenous. That
money supply mechanism (ibid., pp. 32 34). Looking is, central monetary authorities can exert complete
at the data requirements of the model, it seems quite control over the money supply through open market
diYcult to proceed to an empirical estimation. In fact, operations. Second, the analysis has tried to isolate
Miller carried out some comparative static exercises in periods of tight (easy) monetary policy to compare the
order to see the regional eVects of open market opera- performance of regional economies with the national
tions. The conclusions were that open market opera- economy. However both elements are open to ques-
tions were not neutral once the regional dimension tion. The exogenous money supply assumption implies
was introduced, the eVects on each region depending that nancial intermediaries only play a neutral role in
on parameters such as: price; interest rate elasticities of the transmission of monetary changes. There is a grow-
expenditure; money demand functions in each region; ing literature questioning that assumption. Second, it
their relative size in terms of their relative share in the is diYcult to isolate the eVect of monetary policy from
total money stock; value of regional multipliers, etc. the eVect of all other forms of policy ( scal, industrial,
However, what Miller means by monetary non- labour, regional, etc.) being implemented simulta-
neutrality is not that money could aVect regional neously. In other words, the ceteris paribus assumption
output, because his model is one of a fully-employed is unlikely to hold. If valid, these concerns cast doubt
economy, but rather the possibility that monetary on the conclusions reached in the literature surveyed
changes could aVect either prices or interest rates (ibid., in this section.
pp. 72 74, p. 136).
Miller also developed a two-region, reduced-form
Regional impact of open market operations
monetarist model to assess empirically the issue of the
regional impact of US monetary policy. The model The third of the groups in Table 1 is made up of those
was estimated for the period 1969 75 using quarterly papers which have tried to assess the regional impact
data for two regions: the Northeast region and the rest of monetary policy looking at the regional lags which
of the country. The Northeast region grouped the may arise in the process of transmission of open market
reserve districts of Boston, New York, Philadelphia and operations from central to peripheral money markets.
 908 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
This literature is very much tied in with the US speci c to the institutional arrangements of the US,
experience of the 1950s and 1960s regarding the rela- both in terms of the banking system and also in the
tive eVectiveness of open market operations and reserve narrow range of monetary policy instruments consid-
requirements as the Feds instruments of monetary ered. It is not clear therefore how far the results would
control. Those in favour of using open market opera- translate into other contexts.
tions claimed that they were more exible, easily
applied and readily tuned. Those against open market
operations claimed that reserves requirement changes
had a direct eVect on banks in all regional Federal Regional monetary multipliers
Reserve Districts, whereas open market operations Turning from the speci c question of the regional
were transmitted from central to non-central markets impact of monetary policy to the regional nancial
more slowly (see M C P H E T ER S , 1976, p. 1,009). The literature of Table 2, we consider rst the literature on
debate proceeded by testing empirically both the exis- regional money multipliers. The aim of this literature
tence and length of these regional lags in adjusting to is to show how the standard national money multiplier
monetary changes. model is modi ed when introducing the regional
One of the earliest of these papers was S C OT T, dimension, i.e. the eVect that interregional economic
1955. He tried to estimate the lag in the transmission
relationships (trade and nancial ows) may have on
of open market operations from New York to the rest
the regional monetary base. This has been the aim of
of the country. The period of study was 1951:6 to
the works by D OW, 1982, and M O O R E et al., 1985.
1953:5 and the analysis consisted of comparing the
Downloaded by [Fudan University] at 10:40 15 May 2015

We have also included in this category some other

time-pattern of free reserves both by District and by
contributions dealing with the eVect that imbalances
groups of banks, where free reserves was used as an
in the regional balance of payments may have on
inverse index of the eVectiveness of restrictive monetary
regional income (D EI S S , 1978; and G O OD H A RT,
policy. The results indicated that there were important
1989). However, this literature is somewhat tangential
lags in the transmission of open market operations from
to the other literature being surveyed.
central markets to the rest of the country.6 But these
When the money multiplier model is applied in a
conclusions are not supported by other studies (such as
regional setting a new source of base-reserves growth
B RYA N , 1967, and R U F FI N , 1968). To others (such as
M C P H E T ER S , 1976), while these lags may have existed appears in comparison to the national case. In a regional
in the past, they would have disappeared along with setting, apart from open market operations and reserve
the improvements introduced during the 1960s and requirement changes, the regional monetary base may
1970s in the workings of the US money market. also change due to the existence of real ows between
Even if lags were identi ed, they may be due to regions which generate monetary ows, i.e. inter-
other factors than tight monetary policy. Scotts regional exports (imports) either of goods and services
reasoning runs as follows: (1) open market operations or nancial capital. M O O R E et al., 1985, p. 32, develop
determine banks reserves; (2) banks maintain a xed a regional monetary multiplier the value of which
free-reserves ratio; (3) banks transmit tight monetary depends on the proportion of loanable funds invested
policy by reducing lending when their free reserves are in the region rather than outside it.
running out; (4) those banks whose free reserves do D OW, 1982, compared the value of the monetary
not follow the general pattern are therefore isolated multiplier for two regions, one developed and another
from tight monetary policies. But diVerent patterns in underdeveloped, concluding that in the former the
banks free reserves could also be explained, for eVects of a national monetary increase would be more
example, by diVerences in both banks and borrowers expansive due to its lower liquidity preference and
liquidity preference. That is, a bank could have a higher lower propensity to import (less out ows after the
free reserve ratio either because it is unwilling to lend expansion). This model introduced behavioural param-
further or because their customers are unwilling to eters, such as banks and borrowers liquidity prefer-
increase their debt. ence, into the analysis with a role in the transmission of
L AWR E NC E s, 1963, paper was an attempt to dis- monetary changes. In particular, regions experiencing
tinguish between two in uences on regional banks higher liquidity preference would have lower multi-
reserves: (1) changes due to general credit policies; and pliers. The money supply is still treated as exogenous
(2) changes due to banks local economic environment. (in the sense of being reserves-driven), which is incom-
Lawrences study of the US case in the period 1953 61 patible with the current stage of banking development
concluded that banks which primarily serve non-farm in most developed economies.7 On the other hand,
business borrowers and depositors experience the great- this assumption may well be reasonable for regions
est decline in reserve positions in periods of monetary with lower levels of banking development. If peripheral
restraint (ibid., pp. 129 30). In other words, national regions banks are more reserves-constrained than cen-
monetary policy has greatest impact on urban regions. tral regions banks, then the diVerence in capacity for
It must be borne in mind that this literature is very credit growth is exacerbated.
 Regional Finance: A Survey 909
Interregional nancial ows diVerent approaches to the matter of the interregional
interest rate diVerentials.
This literature is mainly concerned more with the
The rst group (such as K E LE H ER , 1979) have
estimation of interregional nancial ows than with
studied the topic from an eYcient markets approach,
their explanation. This may re ect the relative lack
that is, trying to establish a close relationship between
of regional nancial data. However, in addition to
regional and national interest rates. Keleher estimated
presenting the first US data, for the 1930s, H A RT -
a regional interest rate determination model for both
L A ND, 1949, also studied the equilibrating nature of
the mortgage (new and existing homes) and business
interregional nancial ows. Other early studies also
loan (long and short term bank loan by size and
referred to the US (see B OWS H ER et al., 1957, and
revolving credit bank loan by size) markets in the US.
L EE S , 1969); while C A S T EL L S and S I CA RT, 1980, and
(The periods of estimation were 1965:1 to 1977:12 for
B A N CO D E B IL BAO, 1980, concluded similar exercises
mortgages, and 1967:1 to 1976:4 for bank loans.)
for Spain; and S H O RT and N I CH O L A S , 1981, for the
Keleher concluded that regional nancial markets were
UK regions.
integrated, and that interregional diVerentials in interest
In addition money ows between regions have been
rates existed because of regional diVerences in costs,
used to assess intercity relationships. For example,
risks and homogeneity of nancial assets. In fact it
L I EB ER S ON and S C H I WI R I A N , 1962, used interre-
would have been interesting to investigate the nature
gional nancial flows to study intercity relationships in
of these diVerences further, testing for whether the
Iowa, whereas C A R L I N O and L A NG , 1989, tried to
mark-up varies over the business cycle, or varies as
determine economic relationships among US regions
between large and small rms, for example.
Downloaded by [Fudan University] at 10:40 15 May 2015

with nancial transactions taken as a proxy for commer-

The second of the groups we have distinguished is
cial ones.
made up of those papers which have tried to test the
However, closer inspection of the method generally
sensitivity of interregional nancial ows to regional
used to estimate regional nancial variables throws
diVerentials in interest rates. C EB U LA and Z A H A R -
some doubt on the usefulness of the estimations, along
O FF s, 1974, model for US regions for the period
the lines noted above with respect to C O H EN and
1950 71 con rmed the view that regional deposits are
M A ES H I RO, 1977. It is common to estimate regional
insensitive to interest rate diVerentials. However, it is
currency holdings by applying regional GD P shares
suggested that this does not necessarily mean that
to national currency holdings, or personal disposable
markets are segmented but that there are diVerences in
income shares in the case of deposits and credit. While
risks and costs between regions which would remove
this may be the only method available, we need to be
any possibility of pro table transfers of funds between
aware of its implications. First, it is implicitly being
them.
assumed that monetary ows mirror real ones, that is
The third and larger group of papers (H EN D ER S O N ,
that money is mainly demanded for transaction pur-
1944; E DWAR D S , 1964; S CH A A F, 1966; W I NG E R ,
poses. Second, it is also being assumed that no regional
1969; S T R A S Z H EI M , 1971; J A M E S , 1976; O S TAS ,
diVerences exist, in liquidity preference for example.
1977; R O CKO FF, 1977; A S PI N WAL L , 1979; M C K I L -
That is, two regions having the same share in GD P will
L O P and H U T C HI N S O N , 1990; H U T CH I N S O N and
have the same currency, deposits and credit regardless of
M C K I L L O P, 1990; D A MI C O et al., 1990; and F A I NI
the diVerences in liquidity preference which could arise
et al., 1993) have explored factors which might explain
between both.
regional interest diVerentials. Among the factors con-
sidered are the following:

Regional nancial markets ? factors related to market structure such as concentra-

tion ratios, number of institutions, ceiling on interest
Within this block we have distinguished two groups. rates, etc.
The rst deals with the issue of interest rate diVerentials. ? demand factors such as regional pressure on nancial
The other is much more concerned with the issue of resources
regional credit markets and the factors determining ? risk factors, both on the demand (probability of
regional credit availability. default) and supply (risk of bank failure)8
? regional diVerences in transaction costs (cost factors)
Regional interest rate differentials. All of the literature on
? distance from central monetary markets, which may
reduce both the quantity and quality of information
regional interest rate diVerentials, except for H U T C H-
available to agents.
I N S O N and M C K I L L O P, 1990, and F A I N I et al., 1993,
refers to the US, with some papers (such as R O C KOF F, One of the most common explanations given has been
1977) going back to the end of the 19th century. the existence of diVerent costs and risks among regions.
This partly re ects the greater availability of regional H EN D ER S O N , 1944, tested for the ability of these two
nancial data ( particularly interest rates) in the US. factors to explain diVerences in borrowing costs in the
Within the US empirical literature we distinguish three US during the 1920s and 1930s, and concluded that
 910 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
the diVerences were due to costs and not to risk. 1990; M C K I L L O P and H U T C H I NS O N , 1990; DA M -
However, R OC KO FF, 1977, had found that higher I C O et al., 1990; and F A I N I et al., 1993, are among the
interest rates in some US regions during 1870 1914 few which have dealt empirically with the issue of
were due to higher rates of bank failure (a risk factor). regional interest rate diVerentials in countries other
DiVerences in concentration in regional credit than the US. H U T C HI NS O N and M C K I L L O P, 1990,
markets has also been emphasized (by E DWA R D S , 1964, and M C K I L L O P and H U T C H I NS O N , 1990, addressed
and J A M ES , 1976) as a factor explaining interest rate the issue for the UK and Northern Ireland cases,
diVerentials. Edwards found (for the US during 1955 whereas the other two papers addressed the Italian case.
57) a direct relationship between the degree of concen- M C K I L L O P and H U T C HI NS O N , 1990, found some
tration in the local banking market and regional interest evidence regarding the existence of diVerent interest
rates, especially for small and medium sized enterprises rates applied on loans to SMEs. As regards large
(SM Es) (see also J A M ES , 1976). businesses they concluded that no evidence of such
A S PI N WAL Ls, 1979, model was designed to test the regional diVerences existed and that this fact could be
eVect of market concentration on regional mortgage explained by the increased competition among banks
interest rates and was estimated for the rst months of for this segment of the market. As far as personal
1965. The results con rmed the power of variables nancing was concerned, and contrary to what theory
related to market structure and regional income in would have suggested (higher interest rates in isolated
explaining regional interest rates. However, the risk regions), they concluded that: `the interest rates charged
variable seemed not to be signi cant. are, with one or two exceptions, either approximately
Other papers have tried to explain interregional equal across these three regions [England and Wales,
Downloaded by [Fudan University] at 10:40 15 May 2015

diVerentials in interest rates by combining diVerent Scotland and Northern Ireland] or highest in England
factors. This is the case for S C H A A F, 1966, who tried and Wales, ibid., p. 29). However, they also pointed
to explain regional diVerences in mortgage rates in out that these comparisons focusing on interest rates
terms of risk, distance and demand pressure, for the alone omit important aspects of the lending process
period 1964 74. W I NG E R , 1969, criticized Schaaf s relating to bank charges and collateral requirements
use of the loan-value ratio as a measure of risk: `diVerent (ibid.) because these may vary across regions. Once
lenders may not respond alike to the same risk options bank charges and fees were included in total borrowing
because they may diVer in their assets preferences or in costs, some diVerences were found among the three
the regulatory constraints surrounding their operation regions (ibid., pp. 30 31). However, it has to be noted
(W I NG E R , 1969, p. 662). He also pointed out that that the same authors also concluded in another paper
regional diVerences in risk are a consequence of that there is no evidence of a regional constraint nor
regional growth disparities, introducing the possibility of an interest rate structure signi cantly higher for
of circularity. the Northern Ireland nancial sector than that which
Wingers point on the concept and measurement of prevails at the UK level (H U T C HI N S O N and M C K I L -
risk is worth noting because it seems that anything L O P, 1990, p. 430). The evidence would therefore
which cannot be explained by these models is often seem inconclusive. However, it would be interesting to
subsumed under the label `diVerential risk . This, in consider further what is meant by `signi cantly higher
turn, has led to a consideration of regional nancial since sometimes regional diVerences in interest rates
markets (either mortgage or loan markets) as if they are explained as simply re ecting regionally-diVerential
were perfect in the sense that they have been able to risks. As suggested above, consideration would have to
evaluate regional diVerences in costs and risks when be given to how banks measure risk, and/or how far
allocating resources. However, it will be useful to bear risk premia re ect uncertainty rather than risk.
the issue of measuring risk in mind, its relationship to DA M I CO et al., 1990, and F A I NI et al., 1993, also
uncertainty, and whether the latter should also be found some evidence of diVerentials in interest rates
included in the analysis of the regional diVerentials in between the Northern and Southern Italian regions.
interest rates and credit rationing. The model estimated by D A M I C O et al., 1990, p. 209,
Winger also emphasized the importance of regu- concluded that interest rate diVerentials were due to
latory constraints which could aVect lenders behaviour. diVerences in GD P per capita, and to the particular
O S TA S , 1977, for example, re-estimated S C H A A F s, composition of lenders (by size and sector). These two
1966, model in order to include the eVects of state variables accounted for almost 90% of the variability in
usury ceilings on the mortgage market. The model was interregional interest rates in Italy for the period 1969
estimated for 1970 72 and the results showed that 88. Other variables re ecting risk (such as the ratio of
usury ceilings were the most powerful variable in bad loans to total loans) and concentration (such as the
the model.9 S T R A S Z HE I M , 1971, suggested that the Her ndal concentration index) proved to be of minor
existence of segmented credit markets, especially within signi cance.
the personal, mortgage and SME sectors, may account
for diVerences in regional interest rates. Regional credit availability. Although most of the papers
The papers by H U T C H I NS O N and M C K I L L O P, included in this category have a common concern, the
 Regional Finance: A Survey 911
analysis of the factors which determine regional credit
availability, they clearly diVer in approach. The three
approaches could be roughly classi ed in terms of
neo-classical general equilibrium models, the New
Keynesian literature on credit rationing and, nally, the
Post Keynesian literature on regional money and credit.
The rst of these approaches, which involves the
development of some kind of general equilibrium
model, is the least developed approach in this context.
This is so because money and nancial ows are
considered of minor, if any, relevance for regional
economic growth as it is assumed that nancial
resources ow perfectly from one region to another in
order to fund the best investment alternative, i.e. n-
ancial ows simply facilitate the equilibrating mecha-
nism. The papers by M O OR E and N A G U R NE Y, 1989,
and, up to a point, H U G H ES , 1991, 1992, would t in
this category. For example, Moore and Nagurney
assume that, given the regional supply of funds, which
is generally determined by a multiplier process, and the
Downloaded by [Fudan University] at 10:40 15 May 2015

regional demand for funds, which is seen as mainly

depending on interest rates, there will exist some
equilibrium state where `the supply interest rate at a
supply market plus the transaction cost between a pair
of supply and demand markets cannot exceed the
demand interest rate at the demand market, if there is
a positive monetary ow between this pair of supply
and demand markets (M O OR E and N AG UR N EY,
1989, p. 401). Hence, so long as regional credit markets
work properly, equilibrating interregional nancial
ows ensure that monetary variables have no regional
signi cance other than ensuring an eYcient regional
allocation of resources. As a corollary, money only has Fig. 1. Roberts and Fishkinds analysis
regional signi cance in the case of market imperfec- Notes: b1 and b2 are upper and lower bounds for regional
tions. Most research therefore has been concerned with interest rates to uctuate, whereas rj and r* are regional
the identi cation of factors which prevent regional and national interest rates, respectively.
nancial markets working in this way (see R O B ERT S
and F I S H K I ND, 1979; M O O R E and H I L L , 1982;
H A R R I G A N and M C G R EG OR , 1987; and D OW, since it makes comparisons between alternative assets
1987c). diYcult. Third, Roberts and Fishkind also considered
Roberts and Fishkinds paper was inspired by that regional diVerences in liquidity preference and risk
L O S C H s, 1954, study of spatial uctuations in interest aversion could lead to diVerences in terms of interest
rates, and tried to identify the factors which could sensitivity of both supply and demand for assets. Apart
lead to segmentation in regional credit markets. They from their theoretical analysis, Roberts and Fishkind
identi ed three main factors which could explain also quoted some empirical evidence (E B NE R , 1976)
regional segmentation. The rst of them was related to which supported the hypothesis of regional segmenta-
regional agents access to information. Since knowledge tion in financial markets.10
and information about nancial conditions outside the Their overall conclusion may be summarized in
region are only available at some cost, if these costs terms of the following points, which refer to Fig. 1
inhibited pro table nancial arbitrage between regional (see R O B E RT S and F IS HK I N D, 1979, pp. 20 22):
and national nancial markets, then the result would
be regional segmentation in credit markets. Further, ? Regional interest rates may be diVerent from national
since information costs are higher the more isolated rates (either higher or lower).
the region, the most isolated regions are the most likely ? Regional interest rates vary around national ones
to suVer segmentation in their credit markets. Second, between two banks which re ect regional diVerences
the non-homogeneity of nancial assets in terms of in costs.
liquidity, maturity or risk, could be another factor ? The more isolated a region, the wider their bands
explaining segmentation in regional nancial markets, and therefore, the higher is the regional variability
 912 Sheila C. Dow and Carlos J. Rodr guez-Fuentes

Interest A raise in regional demand for credit which is met with financial
rates inflows from outside markets (national market)
Regional market National market S

r2 S

r1

D
D D

Credit

Fig. 2. Moore and Hills analysis

in interest rates. Wider bands may be explained by M O O R E and H I L L , 1982, they incorporated the feed-
the higher transaction costs, lower availability and back eVects of capital ows for the regional bank
higher cost of information on nancial conditions multiplier which Moore and Hill had neglected.
that greater isolation implies. The issue of the regional segmentation of credit
Downloaded by [Fudan University] at 10:40 15 May 2015

? The more isolated the region, the more inelastic markets has also been addressed empirically, by
both supply and demand for regional assets. On M C K I L L O P and H U T C H I NS O N , 1990; H U T C HI N -
the demand side, this inelasticity re ects a higher S O N and M C K I L L O P, 1990; B I A S , 1992; and A M O S ,
dependence by local borrowers on banks as a source 1992. As mentioned earlier, McKillop and Hutchin-
of funds, both because these are likely to be mostly sons (1990) and Hutchinson and McKillops (1990)
households and small businesses, and because of their study of the UK suggested evidence of regional diVer-
isolation from central nancial markets. The higher entials in both interest rates and fees. But they also
inelasticity of supply could re ect higher banks considered whether a credit constraint existed for the
perceived risk or more diYcult risk-assessment for Northern Ireland economy. They compared the assets
peripheral markets (isolated regions). and liabilities of the Northern Ireland clearing banks
? Regional diVerences in IS and IM elasticities would and the British banks for the years 1977, 1980, 1982,
lead to diVerent regional impacts of monetary policy. 1985, and 1987, in order to test whether Northern
M O O R E and H I L L , 1982, added a further explana- Ireland banks had a proportionately lower deposit base,
tion for some kind of regional segmentation within a higher liquidity and lower bank advances, relative to
regional credit markets: the distinction between small the British banks. They concluded that there was no
and large borrowers and lenders within regional evidence of an overall limit on the regional supply
markets. of credit and accordingly, that the Northern Ireland
Moore and Hills analysis, as shown in Fig. 2, con- nancial sector was part of an integrated market despite
siders regional supply of funds as being determined by being distant from the central market, and despite
a regional bank multiplier process and, therefore, lim- particular political and nancial characteristics (H U T -
ited by the regional deposit base as shown by the C H I NS O N and M C K I L L OP, 1990, pp. 428 30).

inelastic portion of the regional supply curve. Demand B I A S , 1992, found some evidence of regional seg-
for funds is considered to operate quite independently mentation in regional US nancial markets. Bias tried
of the regional supply. The excess regional demand for to identify regional diVerences in interest rates sensitiv-
credit at national interest rates could only be met if ity and therefore regional segmentation in nancial
banks lend more locally or borrow from outside the markets for some 12 US states for the period 1967 86.
region. However, they noted that this arbitrage between A M O S , 1992, also found that bank closures in the
local and national markets is less than perfect because US during 1982 88 were closely related to some
some local borrowers and lenders (in the small business regional structural economic patterns and regulations.
and household sectors and, possibly, local banks) do This result was taken as evidence of regional nancial
not have access to national markets (due to mutual lack market segmentation. More recently, A M O S and W I N -
of information). The outcome would be a higher G E ND ER , 1993, oVered further supporting evidence,
regional interest rate, at r2 . in the form of a strong correlation between bank credit
The question of arbitrage between local and national and regional income (at least for 32 out of the 50 states
nancial markets was developed in a more complete considered).
way by H A R R I G A N and M C G R EG OR , 1987. While The second distinctive approach noted above consists
retaining the main ideas about market segmentation of the recent attempts (by S A M O LY K , 1989, 1991,
put forward by R O B ERTS and F I S H K I ND, 1979, and 1994; G R EEN WAL D et al., 1993; and F A I N I et al.,
 Regional Finance: A Survey 913
1993), to extend the New Keynesian credit-rationing (3) the per capita volume of failed business liabilities;
literature to a regional setting. Unlike the modi cations (4) the real growth rate of domestic loans; and (5) the
to neo-classical models discussed above, which intro- ratio of net income to equity capital of commercial
duce information problems as the regional element, the banks. He then tested for diVerences in this relationship
New Keynesians start with an imperfect information for states with high or low non-performing loans ratios,
model, and apply it to regions. This literature focuses respectively. The third model tested for the diVerence
on how asymmetric and imperfect information could between low and high income growth states. Samolyk
lead to low regional capital mobility and, further, to concluded that local banking sector conditions explain
misallocation of nancial resources and regional credit- more of real income growth in states where bank loan
rationing.11 quality has been poor than in those whose banking
The New Keynesian literature points out that, conditions are relatively healthy (S A M O LY K , 1994,
because of the existence of regional segmentation in p. 259), demonstrating the potential power of local
credit markets, local banks wealth, as a determinant of banks to aVect the local economy when regional seg-
banks ability to extend lending, can become one of mentation exists.
the factors which explain regional credit rationing. F A IN I et al., 1993, and M E S S O R I , 1993, have also
Indeed, as local banks are more likely to have superior tried to build up some relationships between, on the
information on local investment opportunities than one hand, local banks monopoly power and banking
outsiders and, therefore, they can monitor them at ineYciencies and, on the other hand, low economic
lower costs, this makes local investors more dependent performance for Southern Italy. In particular, their
on local nancial institutions. analysis suggests that the low productivity of Southern
Downloaded by [Fudan University] at 10:40 15 May 2015

Following on from this analysis, S A M OLYK , 1991, Italy could be explained by ineYciencies in the nancial
1994, developed an empirical model to test for a sector, due to informational problems in Southern
relationship between banking conditions and economic credit markets.
performance at state level in the US. The underlying The New Keynesian literature, as summarized in
hypothesis was that the existence of information costs Fig. 3, suggests that asymmetric information explains
may lead to credit constraints in some nancially dis- credit rationing since it inhibits the provision of credit
tressed regions, but not in nancially sound regions. by national (outsider) institutions in regional markets
Samolyk tested (for the period 1983 90) the relation- when local ones fail to provide. This conclusion poses
ship between state income growth rates and regional the further question of how far branching of national
credit conditions such as: (1) the real growth rate of banks in credit-constrained regions would get round
loan loss reserves; (2) the non-performing loan share; the problem of asymmetric information. P O RT E O U S ,

Local banks:
better information
lower monitoring costs
market power

REGIONAL Credit REGIONAL

DEMAND rationing SUPPLY

Outside banks:
imperfect information
higher monitoring costs
unwillingness to lend regionally

Outcome:
lower interregional capital mobility
regional demand strongly depends on local demand
local banks inefficiencies are passed on to local demand for credit

Fig. 3. New Keynesian theory

 914 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
1995, concludes that the optimal banking structure to nancial institutional structure, in particular the ques-
address the problem of monitoring costs for rms in tion of how far branch banking would alter the conclu-
peripheral regions is a mixed one, with small local sions of the US literature referring to unit banking. It
banks providing credit for small local rms. might be concluded that regions with a branch bank
Rather than focusing on why a perfect ow of nan- system face a horizontal supply of funds because
cial resources among regions does not exist, the Post regional bank branches are able to lend beyond their
Keynesian literature on regional money and credit takes regional-deposit base. This, in turn, means that credit
market imperfection as the norm. Instead it focuses on availability is no longer a problem for regions so long
the study of regional patterns of credit creation, and how as its supply is perfectly elastic at some mark-up over
these may vary from one region to another. In so doing, the national interest rate. However, Dow concluded
Post Keynesian theory makes use of both Chicks stages that even when a branch banking system exists, we
of banking development and the Keynesian principle of cannot assume a perfectly elastic regional supply of
liquidity preference (see C H I C K and D OW, 1988; D OW, funds. Although a low regional deposit base would not
1987a, 1988, for a theoretical account; D OW, 1987b, for necessarily mean less regional credit, it might mean a
an application to an open economy; and D OW, 1990, higher regional liquidity preference on behalf of the
1992, and C H IC K , 1993, for applications to the Canad- national banks and this, in turn, will be the factor
ian, Scottish and EC cases, respectively). which will constrain the regional extension of bank
Although there are some parallels between the New lending. Therefore, despite the endogenous character
Keynesian and the Post Keynesian theories, a closer of the regional money supply, credit availability still
look reveals signi cant diVerences. A particular feature remains as a variable that matters for regional analysis.
Downloaded by [Fudan University] at 10:40 15 May 2015

of the Post Keynesian theory is that its analysis addresses Hence, regional credit rationing is not seen as a
both the supply-side and the demand-side of the unicausal situation explained by a regional-
regional credit market. New Keynesian literature is discriminatory behaviour on the part of the nancial
mainly concerned with the supply-side issue of how system (mainly banks) which, in turn, leads to an
imperfect information segments regional markets. uneven regional distribution of credit, but as a multi-
Regional credit rationing could arise then as a result causal situation in which all sectors in the region are
of the unwillingness of non-local nancial institutions involved (see Fig. 4). This is what D OW, 1992, has
to lend within the region (because of their lack of labelled `defensive nancial behaviour, which we now
information to assess local project riskiness and pro t- explore in more detail in terms of the regional supply
ability). However, Post Keynesians point out that credit of and demand for credit.
rationing could also be explained by demand factors to Regarding the supply side, Post Keynesian theory
the extent that the amount of regional credit is the considers that regional credit supply is aVected both by
result of the interaction between supply and demand, regional liquidity preference and the stage of banking
and because both functions are interdependent, being development. The stage of banking development deter-
aVected by changes in liquidity preference. mines banks ability to extend credit regardless of their
This is evident in D OW s, 1987c, modi cation of deposit base, either regional or national, i.e. the degree
the M O O R E and H I L L , 1982, analysis by means of the of endogeneity of money supply. The lower the stage
addition to Moore and Hills analysis of three further of bank development, the more applicable the money
considerations. First, the demand for and supply of multiplier model is. This would imply that regions
credit may be interdependent. Dow suggested that, having banking systems in lower stages of development
apart from open market operations and interregional would be more constrained by, say, low saving or
nancial ows, the regional monetary base could also deposit ratios than others. Thus, local banks at an
be in uenced by the regional demand for credit. The earlier stage of development than national or inter-
second factor was the recognition of the speculative national banks will nd themselves at a competitive
component in the demand for money and, hence, the disadvantage, including a lesser capacity to create credit.
role played by liquidity preference in the regional Post Keynesian theory allows liquidity preference to
credit creation process. The introduction of both the aVect regional suppliers of credit and regional demand
endogeneity of money and liquidity preference led to for credit. From the banks point of view, liquidity
a reversal of Moore and Hills causation. That is, instead preference in uences the willingness to lend within the
of only considering the possibility that changes in region when regional perceived-risk is higher or its
regional income could lead to changes in regional assessment is more diYcult.12 The New Keynesian
deposits and credits, as Moore and Hill assumed, Dow approach refers to risk assessment as an objective pro-
also opened the possibility for changes in regional cess, where in principle full information would allow
liquidity preference (due to greater/lower con dence the generation of a risk measure. The Post Keynesian
in the regional economy) leading to endogenous approach rather sees all credit-risk assessment as being
changes in regional credit and, therefore, changes in subject to uncertainty of varying degrees; that uncer-
regional income. tainty is perceived to be greater the more remote the
Third, D OW, 1987c, focused on the signi cance of borrower from the lender (where remoteness may be
 Regional Finance: A Survey 915

Local investors: REGIONAL DEMAND

willingness to borrow FOR CREDIT

LIQUIDITY Local savers:

PREFERENCE preferences in terms of
risk, maturity and profitability

Banks:
willingness to lend within
the region
Downloaded by [Fudan University] at 10:40 15 May 2015

Banks: REGIONAL SUPPLY

STAGE OF BANKING ability to lend
DEVELOPMENT OF CREDIT
beyond their deposit base

Fig. 4. Post Keynesian theory

spatial, cultural, etc.: see P O RT E O U S , 1995). Banks which depends on its stage of development), it could
liquidity preference may hence be in uenced by both also increase the regional supply of credit. It is this
regional expectations regarding regional income interdependence between supply of and demand for
growth, regional instability, etc., and by the expected credit which makes it diYcult to identify whether any
eVects of monetary conditions created by central bank. regional credit-gap exists. Some authors have tried to
Liquidity preference not only aVects lenders behav- assess the matter in terms of declining long-run credit
iour but also exerts its eVect on savers behaviour. For shares by peripheral regions and have found no clear
example, higher liquidity preference will encourage evidence of such a trend. P O RT E O US , 1995, for
savers to adopt more liquid portfolios, where that example, presents some empirical evidence for Australia
liquidity is more likely to be supplied by extra-regional (1950 84) and Canada (1975 90), suggesting that,
assets. An increase in liquidity preference by regional overall `there is little strong evidence of systematic
agents in peripheral regions could then lead to an discrimination by national banks as seen in long-run
out ow of nancial resources to central regions which credit shares or in the extent of rationing (P O RT EO U S ,
may reduce local availability of funds. Whether or 1995, p. 193). Porteous further concluded that: (1)
not this out ow aVects regional credit availability will wealthier regions have higher credit shares; (2) there is
depend on: (1) the ability of the banking sector to no evidence suggesting the existence of a long-run
expand credit regardless of its regional deposit base; decreasing trend in credit shares for peripheral regions;
and (2) the eVect that such regional out ows have on (3) national monetary policy does not discriminate
banks own regional liquidity preference. However, against peripheral regions; and (4) some evidence sug-
what is worth noting in this process is that this eVect gests the existence of credit rationing in peripheral
has its origins in other sectors than the nancial one. regions, but this evidence is not conclusive.
With regard to the demand side, we have to take However, Post Keynesian theory does not claim
into account the eVect that liquidity preference could long-run decreases in regional credit shares by peri-
have on the regional demand for credit. For example, pheral regions. Rather it claims unstable patterns in
lower expectations regarding the regional economy regional credit creation in the sense that credit creation
(higher liquidity preference) could lower the regional can fuel expansions and enhance recessions generating
demand for funds to the extent that investors are less greater instability.13 Since money is credit-driven, rather
willing to run into debt. Higher regional expectations than the other way round, the issue is no longer limited
could drive up regional demand for credit and, to the to looking at whether banks lend more than they
extent that the banking system shares the optimism and borrow regionally, as is often suggested. The matter is
is able to extend credit beyond its deposit-base (a factor no longer how a xed amount of credit is divided up
 916 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
among regions, but how credit is created (or not) in a more general sense, as Post Keynesian monetary
regionally. The focus then is on the interdependent theory suggests, both the theoretical and empirical
relationships between credit, deposits, money supply focus of regional analysts would be diVerent from that
and income over cycles. For example, credit is expected in most of the literature reviewed. In particular, the
to increase in economic upturns because of higher focus would shift to concentrate on the institutional
banks and borrowers willingness to lend and borrow, structure of banking, the basis for banks risk assessment
respectively. The deposit total depends on whether and regional diVerences in nancial behaviour. This
most of this credit ows outside the region in terms of un nished business means that there is still a rich
imports and capital out ows. Therefore, a high credit/ research agenda for regional nance.
deposit ratio is expected during expansions in peri-
pheral regions. However, what happens in recessions is
more diYcult to predict. During downturns, credit NOT E S
demand and supply are expected to be low because of 1. This seems to apply to B E AR E s, 1976; M AT H U R and
higher liquidity preference on behalf of both investors S TE IN s, 1980; and G AR RI SO N and K O RTS , 1983,
and lenders. However, a regional credit gap could still papers, since they are regional applications of A ND-
exist to the extent that there is a need to nance E RSEN and J O R DA N s, 1968, paper. As we shall see, the
working capital. At the same time, higher liquidity only diVerence between these papers is that, for example,
preference could lead to nancial out ows if safer and B E AR E , 1976, developed his analysis for some Canadian
regions instead of for the whole Canadian economy.
more liquid nancial assets are not provided within the
The same point could be made of both M AT H U R and
region. Whether the credit/deposit ratio remains high
Downloaded by [Fudan University] at 10:40 15 May 2015

S TE IN , 1980 and G AR RI SO N and K O RT, 1983, since

or low with regard to the expansion phase will depend they also applied a reduced-form model to some eight
on the signi cance of each of the factors above men- U S regions instead of to the whole U S economy.
tioned (credit demand, credit supply and nancial out- 2. Most of this literature deals with the low eVectiveness of
ows). However, no deterministic relationships are to monetary and exchange rate policies in small open eco-
be expected because some of the factors aVecting them nomies. The underlying argument is that exogenous and
cannot be deterministically measured or foreseen (such expansive monetary shocks will lead to imbalances in the
as changes in liquidity preference), or can be modi ed external sector due to the highly open character of these
by others (such as the institutional setting). economies (see, for example, A L LY, 1975; K H AT K H AT E
and S H O RT, 1980; C O R BO and O SS A , 1982; C A RA M ,
1985, 1993; C RUS O L , 1986; B L EJ ER , 1988; and
C O N C L US I O N S W O RR EL , 1991. However, some authors have also
recognized some eVectiveness for certain policies such as
In this survey of the regional nance literature, the selective credit controls (K H AT KH AT E and V IL-
regional impact of monetary policy has been an import- L A N U EVA , 1978; C RUSO L , 1986; and B L EJ ER , 1988).
ant focus. We have tried to show that most of this 3. One issue which remains unexplained within the monet-
literature consists, to a large extent, simply of regional arist position is the one regarding how the economy is
extensions to the national discussion regarding how able to grow in the short-run whereas real factors, the
monetary policy aVects economic activity. Empirical only ones which can cause growth according to this view,
work has therefore been focused on testing whether are supposed to remain xed within this period of time.
money causes regional business cycles as well, or 4. Early models developed by, among others, C Z AM A NS KI ,
whether national monetary conditions aVect regions 1969, and G L I CKM A N , 1977, would be included in this
group, and clearly contrast with those other neoclassical
diVerently because of structural diVerences.
models of regional growth which explicitly have denied
Since regions are very small open economies and do any role for monetary variables to play in the explanation
not have monetary tools at their disposal, most of of regional economic growth. An example of the latter
the analysis regarding the potential regional eVect of can be found in B O RTS , 1968.
monetary policy has been developed within the stan- 5. The indicator of degree of monetary tightness was the rate
dard open-economy version of the IS LM model. of growth of M2 (money stock plus net time deposits).
Special attention has been given either to identifying 6. However, there have also been contributions which have
regional structural diVerences in terms of IS and LM arrived at the conclusion that such lags do not exist. See,
slopes, or to those factors which may lead to some for example, B RYA N , 1967. Some others have also
segmentation in regional credit markets. In fact, these argued that even if they existed, lags would be reduced
two approaches along with those which have consid- by improvements introduced in the workings of the
monetary markets during the 1960s and 1970s. However,
ered money as being the cause of regional business
M C P H ET ERS , 1976, found no signi cant change
cycles, have attracted most empirical eVort. between the periods 1946 60 and 1961 69.
Despite other methodological diVerences, all 7. We are referring here to Chicks model of stages of
approaches but the Post Keynesian one share a common banking development (see C H I CK , 1986). See also
assumption, that the supply of money is exogenous to C H I CK and D OW, 1988; and C H IC K , 1993, for two
the national economic system but endogenous in a applications to a regional setting.
regional setting. However, if money were endogenous 8. R O CKOF F, 1977, has associated the high interest rates
 Regional Finance: A Survey 917
existing in some U S regions during 1870 1914 with 1989; B ERN A NK E , 1993; G RE EN WA L D and S TI G L IT Z ,
their higher rate of bank failure. 1993; and G E RTL ER and G ILCH RIS T, 1993. C OS CI ,
9. Another diVerence with Schaaf s ndings was that dis- 1993, oVers a survey of this literature.
tance was not signi cant any more. 12. This is the New Keynesian imperfect-information
10. The empirical evidence given by E B NE R , 1976, showed argument.
that some regional diVerences existed in terms of the 13. See D OW, 1993, chapter 3, speci cally pp. 38 40.
sensitivity of savings to changes in national interest rates. M I N SKY, 1982, chapter 5, oVers a comprehensive
11. On this point see, among others, the papers by S T I GL I T Z account of the relationship between lending expansion
and W E IS S, 1981; G ERTL ER , 1988; B ERN A NK E and (contraction) and business cycles.
B L IN D ER , 1988, 1992; B ER NA N KE and G ERTLE R ,

RE F E RE N C E S
A L L EN L. and P RI C E D. (1984) Variations on the ow of credit from regional nancial institutions as a source of instability in
the regional economy: an application of causal analysis, Reg. Sci. Perspectives 14, 3 14.
A L LY A. (1975) The potential for autonomous monetary policy in small developing countries, in S E L DWYN P. (Ed) Development
Policy in Small Countries. Institute of Development Studies, University of Sussex.
A LVA REZ L L A NO R. and A NDR EU J. M. (1978) Flujos nancieros regionales y banca: un analisis comparativo international,
Bolet n de Estudios Economicos 103, 131 65.
Downloaded by [Fudan University] at 10:40 15 May 2015

A M O S O. M. (1992) The regional distribution of bank closings in the United States from 1982 to 1988, Southern Econ. J. 58,
805 15.
A M O S O. M. and W IN G END ER J. R. (1993) A model of the interaction between regional nancial markets and regional
growth, Reg. Sci. Urban Econ. 23, 85 110.
A N D ERS EN L. C. and J O R DA N J. L. (1968) Monetary and scal actions: a test of their relative importance in economic
stabilisation, Fed. Res. Bank St Louis Rev. 50, 11 21.
A SP IN WA L L R. C. (1979) Market structure and commercial bank mortgage interest rates, South. Econ. J. 36, 376 84.
B A NC O D E B I L BAO (1980) Flujos monetarios regionales y balanza de pagos, Situacion 7, 18 49.
B A RTH J., P H AU P M. and P I ERC E D. M. (1975) Regional impact of open market operations on member bank reserves, J.
Econ. Bus. 28(1).
B E AR E J. B. (1976) A monetarist model of regional business cycles, J. Reg. Sci. 16, 57 63.
B E RN AN KE B. S. (1993) Credit in the macroeconomy, Fed. Res. Bank N Y Quart. Rev. 18, 50 70.
B E RN AN KE B. S. and B L IN D E R A. S. (1988) Is it money or credit, or both, or neither?, Am. Econ. Rev. 78, 435 39.
B E RN AN KE B. S. and B L IN D ER A. S. (1992) The federal funds rate and the channels of monetary transmission, Am. Econ.
Rev. 82, 901 21.
B E RN AN KE B. S. and G ERTL ER M. (1989) Agency costs, net worth, and business uctuations, Am. Econ. Rev. 79, 14 31.
B I A S P. V. (1992) Regional nancial segmentation in the United States, J. Reg. Sci. 32, 321 34.
B L A KE N. (1995) The regional implication of macroeconomic policy, Oxf. Econ. Pap. 11, 145 64.
B L E JE R M. I. (1988) Growth, investment, and the speci c role of scal policies in very small developing economies, in J O RG E
A. and S L AZ A R- C A RR ILL O J. (Eds) Foreign Investment, Debt and Economic Growth in Latin America. Macmillan, London.
B O RTS G. H. (1968) Regional economic models: growth and capital movements among US regions in the post-war period,
Am. Econ. Rev. 58, 155 61.
B OWSH ER N. N., D A AN E J. D. and E IN Z IG R. (1957) The ow of funds between regions of the United States, Pap. Reg. Sci.
Ass. 3, 139 59.
B OY D W. E. (1970) Regional response to monetary policy 1960 1969, unpublished Ph.D. dissertation, Florida State University.
B RYAN W. R. (1967) Bank adjustment to monetary policy: alternative estimates of the lag, Am. Econ. J. 57, 855 64.
C A RA M A. R. (1985) Guidelines for monetary policy in small developing countries, K AM I RA M IDE S J ., B RI G U L IO L . and
H O O G EN DA R K H . N . (Eds) The Economic Development of Small Countries: Problems, Strategies and Policies. Eburon, Delft.
C A RA M A. R. (1993) The repercussions of nancial imbalances in Surinam, World Develop. 21, 291 99.
C A RL I NO G. and L A N G R. (1989) Interregional ows of funds as a measure of economic integration in the Unites States, J.
Urban Econ. 26, 20 29.
C A RR H. C. (1960) A note on regional diVerences in discount rates, J. Finance 15, 62 68.
C A ST EL L S A. and S I CA RT F. (1980) Flujos nancieros interregionales: una aproximacion, Hacienda Publica Espanola 63, 43 96.
C EB U L A R. J. and Z A H A ROFF M. (1974) Interregional capital transfers and interest rate diVerentials: an empirical note, Ann.
Reg. Sci. 8, 87 94.
C H A SE E CO N O ME TR IC A SS O CI AT ES I N C (1981) Rural impacts of monetary policy, Agr. Econ. Res. 33, 1 11.
C H I CK V. (1986) The evolution of the banking system and the theory of saving, investment and interest, Economies et Societes
20, Monnaie et Production 3, 111 26.
C H I CK V. (1993) Some scenarios for money and banking in the EC and their regional implications, in R I M A I. H. (Ed) The
Political Economy of Global Restructuring, vol. 2, Trade and Finance. Edward Elgar, Cheltenham.
C H I CK V. and D OW S. C. (1988) A post Keynesian perspective on the relation between banking and regional development,
in A RES TI S P. (Ed) Post Keynesian Monetary Economics: New Approaches to Financial Modelling. Edward Elgar, Cheltenham.
 918 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
C H I CK V. and D OW S. C. (1997) Competition and the future of the European banking and nancial system, in C O H E N A.
J., H AG ERM A NN H. and S M I TH I N J. (Eds) Money, Financial Institutions and Macroeconomics, pp. 253 70. Kluwer, Boston.
C O H EN J. and M A ESH I RO A. (1977) The signi cance of money on the state level, J. Money, Credit & Banking 9, 672 78.
C O RB O V. and O SSA F. (1982) Econom as pequen as, y abiertas: una vision general, Working Paper No. 80, Ponti cia
Universidad Catolica de Chile.
C O SCI S. (1993) Credit Rationing and Asymmetric Information. Avebury, Aldershot.
C RUS O L J. (1986) La politique monetaire dans les tres petites economies insulaires: lexperience des iles de la caraibe
anglophone: Barbade, Jamaique, Trinidad et Tobago, Mondes en Developpement 14, 109 22.
C Z AM A NS KI S. (1969) Regional econometric models: a case study of Nova Scotia, in S CO TT A. J. (Ed) Studies in Regional
Science, London Papers in Regional Science. Pion, London.
D A M I CO N., P AR I GI G. and T R IF ILI D I S M. (1990) I tassi dinteresse e la rischiosita degli impieghi bancari, in B AN CA
DI TA L I A (Ed) Il Sistema Finnaziario nel Mezzogiorno, special issue of the Contributi allanalisi economica, pp. 305 347. Banca
dItalia, Rome.
D AV I S R. and B A N KS L. (1965) Interregional interest rate diVerentials, Fed. Res. Bank NY Monthly Rev., August, pp. 165 74.
D E ISS J. (1978) The regional adjustment process and regional monetary policy, Rivista Internazionale di Scienzi Echonomiche e
Commerciali 25, 858 77.
D OW S. C. (1982) The regional composition of the money multiplier process, Scot. J. Pol. Econ. 29, 22 44.
D OW S. C. (1987a) Money and regional development, Studies Pol. Econ. 23, 73 94.
D OW S. C. (1987b) Post Keynesian monetary theory for an open economy, J. Post Keynesian Econ. 9, 237 57.
D OW S. C. (1987c) The treatment of money in regional economics, J. Reg. Sci. 27, 13 24.
D OW S. C. (1988) Incorporating money in regional economic models, in H A RR IG A N F. and M C G REG O R P. (Eds) Recent
Downloaded by [Fudan University] at 10:40 15 May 2015

Advances in Regional Economic Modelling, Papers in Regional Science 19. Pion, London.
D OW S. C. (1990) Financial Markets and Regional Economic Development: The Canadian Experience. Avebury, Aldershot.
D OW S. C. (1992) The regional nancial sector: a Scottish case study, Reg. Studies 26, 619 31.
D OW S. C. (1993) Money and the Economic Process. Edward Elgar, Cheltenham.
D OW S. C. (1996) European monetary integration, endogenous credit creation and regional economic development, in
V EN CE - D EZ A X. and M ET CA L FE J. F. (Eds) Wealth from Diversity. Kluwer, Dordrecht.
D R EESE G. R. (1974) Banks and regional economic development, South. Econ. J. 40, 647 66.
E BN ER M. (1976) Development, estimation and forecasting accuracy of regional nancial models: an application within the
State of Florida, unpublished Ph.D. dissertation, University of Florida, Gainesville.
E DWARD S F. (1964) Concentration in banking and its eVects on business loan rates, Rev. Econ. Statist. 46, 294 300.
F AI N I R., G AL L I G. and G IA N NI N I C. (1993) Finance and development: the case of Southern Italy, in G I OVANN IN I A. (Ed)
Finance and Development: Issues and Experience. Cambridge University Press, Cambridge.
F E RN AN D EZ F. and A N D RE U J. M. (1978) Algunas consideraciones sobre los ujos nancieros geogra cos, Estudios Regionales
2, 119 33.
F I SH KI N D H. H. (1977) The regional impact of monetary policy: an economic simulation study of Indiana 1958 1973, J. Reg.
Sci. 17, 77 88.
F R IED MA N M. and M EI SEL M A N D. (1963) The relative stability of monetary velocity and the investment multiplier in the
United States, in C O M MI SS IO N O F M O N EY A N D C RED I T, Impacts of Monetary Policy. Prentice-Hall, Englewood CliVs, NJ.
G A R RIS O N C. B. and C H A NG H. S. (1979) The eVect of monetary and scal policies on regional business cycles, Int. Reg.
Sci. Rev. 4, 167 80.
G A R RIS O N C. B. and K O RT J. R. (1983) Regional impact of monetary and scal policy: a comment, J. Reg. Sci. 23, 249 61.
G E RTL ER M. (1988) Financial structure and aggregate economic activity: an overview, J. Money, Credit & Banking 20, 559 88.
G E RTL ER M. and G I L CH RI ST S. (1993) The role of credit market imperfections in the monetary transmission mechanism:
arguments and evidence, Scand. J. Econ. 95, 43 64.
G I L B ERT J. C. (1937 1938) The mechanism of interregional redistribution of money, Rev. Econ. Studies 5, 187 94.
G L ICK MA N N. J. (1977) Econometric Analysis of Regional Systems: Explorations in Model Building and Policy Analysis. Academic
Press, New York.
G L ICK MA N N. J. (1980a) The Urban Impacts of Federal Policies. The Johns Hopkins University Press, Baltimore.
G L ICK MA N N. J. (1980b) Impact analysis with regional econometric models, in P L E ET ER S. (Ed) Economic Impact Analysis:
Methodology and Applications. Studies in Applied Regional Science 19. Martinus NijhoV, Amsterdam.
G O O D H ART C. A. E. (1989) Money, Information and Uncertainty, 2nd edition. Macmillan, London.
G R EEN WA L D B. C., L EV I NS O N A. and S T IG L I TZ J. E. (1993) Capital market imperfections and regional economic
development, in G I OVANN I NI A. (Ed) Finance and Development: Issues and Experience. Cambridge University Press,
Cambridge.
G R EEN WA L D B. C. and S T IG L IT Z J. E. (1993) Financial market imperfections and business cycles, Quart. J. Econ. 58, 77 114.
H AR RI GA N F. J. and M C G RE GO R P. G. (1987) Interregional arbitrage and the supply of loanable funds: a model of
intermediate nancial capital mobility, J. Reg. Sci. 27, 357 67.
H ARTL AN D P. C. (1949) Interregional payments compared with international payments, Quart. J. Econ. 63, 392 407.
H END ERS H OT T P. and K IDWE L L D. (1978) The impact of relative security supplies: a test with data from a regional tax-
exempt bond market, J. Money, Credit & Banking 20, 337 47.
H END ERS O N J. (1944) Regional diVerentials in interest rates, South. Econ. J. 11, 113 32.
 Regional Finance: A Survey 919
H U GH ES M. (1991) General equilibrium of a regional economy with a nancial sector, part I: an accounting framework with
budget and balance sheet linkages, J. Reg. Sci. 31, 385 96.
H U GH ES M. (1992) General equilibrium of a regional economy with a nancial sector, part II: a simple behavioural model, J.
Reg. Sci. 32, 19 37.
H U TCH IN SO N R. W. and M C K I L L O P D. G. (1990) Regional nancial sector models: an application to the Northern Ireland
nancial sector, Reg. Studies 24, 421 31.
J A ME S J. (1976) Banking market structure, risk and the pattern of local interest rates in the US, 1893 1911, Rev. Econ. Statist.
58, 453 62.
K A NN A N R. (1987) Banking development and regional disparities, Indian Econ. J. 35, 58 76.
K EL EH E R R. E. (1977) Fundamental determinants of credit volume: a survey and regional application, technical papers,
Federal Reserve Bank of Atlanta, Atlanta, GA.
K EL EH E R R. E. (1979) Regional credit market integration: a survey and empirical examination, Technical Papers, Federal
Reserve Bank of Atlanta, Atlanta, GA.
K H AT KH ATE D. R. and S H O RT B. K. (1980) Monetary and central banking problems of mini states, World Develop. 8,
1,017 25.
K H AT KH ATE D. R. and V I L L A NU E VA D. P. (1978) Operation of selective credit policies in less developed countries: certain
critical issues, World Develop. 6, 979 90.
K O ZLOWSKI P. J. (1991) Integrating money into regional models of leading indicators, Rev. Reg. Studies 21, 235 48.
L AW R ENC E R. J. (1963) The regional impact of monetary policy, unpublished Ph.D. dissertation, University of Michigan.
L EE S F. A. (1969) Interregional ows of funds through state and local government securities (1957 1962), J. Reg. Sci. 9, 79 86.
L I EBE RSO N S. and S CH W I RI AN K. P. (1962) Banking functions as an index of inter-city relations, J. Reg. Sci. 4, 69 81.
Downloaded by [Fudan University] at 10:40 15 May 2015

L O SCH A. (1954) The Economics of Location. Yale University Press, New Haven, CT.
M ATHU R V. K. and S TE IN S. (1980) Regional impact of monetary and scal policy: an investigation into the reduced form
approach, J. Reg. Sci. 20, 343 51.
M ATHU R V. K. and S T EI N S. (1982) The regional impact of monetary and scal policy: some further results, Pap. Reg. Sci.
Ass. 50, 67 74.
M ATHU R V. K. and S T EI N S. (1983) Regional impact of monetary and scal policy: a reply, J. Reg. Sci. 23, 263 65.
M C K I L L O P D. G. and H U T CH IN SO N R. W. (1990) Regional Financial Sectors in the British Isles. Avebury, Aldershot.
M C P H ET ER S L. R. (1976) Banking system diVusion of open market operations, Southern Econ. J. 43, 1,009 16.
M E SSO RI M. (1993) Banking and nance in the Italian Mezzogiorno: issues and problems, paper presented at the International
Congress on The European Periphery Facing the New Century, Santiago de Compostela.
M E YE R P. A. (1967) Price discrimination, regional loan rates and the structure of the banking industry, J. Finance 22, 37 48.
M I L L ER R. J. (1978) The Regional Impact of the Monetary Policy in the United States. Lexington Books, Lexington, MA.
M I N NS R. and M ARTI N R. (1995) Undermining the nancial basis of regions: the spatial structure and implications of the
U K pension fund system, Reg. Studies 29, 125 44.
M I N SKY H. P. (1982) In ation, Recession and Economic Policy. Wheatsheaf, Brighton.
M O O R E C. L. (1979) Banking and the regional income multiplier, Northeast Reg. Sci. Rev. 9, 1 7.
M O O R E C. L. and H I L L J. M. (1982) Interregional arbitrage and the supply of loanable funds, J. Reg. Sci. 22, 499 512.
M O O R E C. L., K A R ASK A G. J. and H ILL J. M. (1985) The impact of the banking system on regional analyses, Reg. Studies
19, 29 35.
M O O R E C. L. and N AG U RN EY A. (1989) A general equilibrium model of interregional monetary ows, Environ. Plann. A,
21, 397 404.
O STAS J. R. (1977) Regional diVerences in mortgage costs: a re-examination, J. Finance 32, 1,774 78.
P ETE RSO N M. (1973) Some evidence on intra-regional diVerences in yields and costs of mortgage lending, Land Econ. 49,
96 99.
P ORTEO U S D. J. (1995) The Geography of Finance: Spatial Dimensions of Intermediary Behaviour. Avebury, Aldershot.
R I CH AR D SO N H. W. (1972) Regional Economics: Location Theory, Urban Structure and Regional Change. World University, London.
R I CH AR D SO N H. W. (1973) Regional Growth Theory. Macmillan, London.
R O B ERTS R. B. and F I SH KI ND H. (1979) The role of monetary forces in regional economic activity: an econometric
simulation analysis, J. Reg. Sci. 19, 15 29.
R O CKOF F H. (1977) Regional interest rates and bank failures, 1870 1914, Explorations in Econ. Hist. 14, 90 95.
R O D R I G U EZ - F U E NT ES C. J. (1993) Dinero, pol tica monetaria y econom a regional: una aproximacion, Situacion 2, 27 43.
R O D R I G U EZ - F U E NT ES C. J. (1995) Pol tica monetaria y econom a regional, unpublished Ph.D. thesis, Department of Applied
Economics, University of La Laguna, Spain.
R U FF IN R. J. (1968) An econometric model of the impact of open market operation on various bank classes, J. Finance 23,
625 37.
S AM O LY K K. A. (1989) The role of banks in in uencing regional ows of funds, Working Paper 8914, Federal Reserve Bank
of Cleveland.
S AM O LY K K. A. (1991) A regional perspective on the credit view, Fed. Res. Bank Cleveland, Econ. Rev. 27, 27 38.
S AM O LY K K. A. (1994) Banking conditions and regional economic performance: evidence of a regional credit channel, J.
Monetary Econ. 34, 259 78.
S COT T I. O. (1955) The regional impact of monetary policy, Quart. J. Econ. 69, 269 84.
S CH A AF A. H. (1966) Regional diVerences in mortgage nancing costs, J. Finance 21, 84 94.
 920 Sheila C. Dow and Carlos J. Rodr guez-Fuentes
S H O RT J. and N I CH O L A S D. J. (1981) Money Flows in the UK Regions. Gower, Farnborough.
S TI GL IT Z J. E. and W EI SS A. (1981) Credit rationing in markets with imperfect information, Am. Econ. Rev. 71, 393 410.
S TR AS ZH EIM M. R. (1971) An introduction and overview of regional money and capital markets, in K AI N J. F. and M EY ER
J. (Eds) Essays in Regional Economics. Harvard University Press, Cambridge, M A.
T H U RS TO N T. B. (1976) Regional interaction and the reserve adjustment lag with the commercial banking sector, J. Finance
31, 1,443 56.
T O AL W. D. (1977) Regional impacts of monetary and scal policies in the post war period: some initial tests, Technical Paper,
Federal Reserve Bank of Atlanta, Georgia.
W IN G ER A. (1969) Regional growth disparities and the mortgate market, J. Finance 24, 659 62.
W O RR EL D. (1991) Fiscal and monetary policies in small economies, in Y I N- K A N N W E N and S EN G U PTA Y. (Eds) Increasing
the International Competitiveness of Exports from Caribbean Countries. Economic Institute of the World Bank, Washington, DC.
Downloaded by [Fudan University] at 10:40 15 May 2015