1

BANK-LENDING CHANNEL AND NON-FINANCIAL FIRMS: EVIDENCE FOR SPAIN

Santiago Carbó Valverde♣

(Universidad de Granada and Federal Reserve Bank of Chicago)

Departamento de Teoría e Historia Económica.

Facultad de Ciencias Económicas y Empresariales.

Universidad de Granada. Campus Universitario de Cartuja s/n.

E-18071 Granada

Tel: +34 958243717

E-mail: scarbo@ugr.es

Rafael López del Paso

Departamento de Teoría e Historia Económica.

Facultad de Ciencias Económicas y Empresariales.

Universidad de Granada. Campus Universitario de Cartuja s/n.

E-18071 Granada

Tel: +34 958243717

E-mail: lopezdelpaso@yahoo.es

♣ Corresponding author

Acknowledgements: Financial support from MEC-FEDER, SEJ2005-04927 are acknowledged and appreciated

by the authors. Santiago Carbó also acknowledges financial support from Junta de Andalucia, SEJ 693

(Excellence Groups). The views in this paper are those of the authors and may not represent the views of the

Federal Reserve Bank of Chicago or the Federal Reserve System.

2

Abstract: During the 1990s, liquidity was relatively abundant in the European Union and the

European central banks mostly developed a relaxed monetary policy. While the bank lending channel

view of the monetary policy would have suggested an increase in loans to firms in this context, the

demand for bank corporate lending, however, slowed down, suggesting that monetary policy was not

effective in this area. This article analyses how the financing behaviour of Spanish firms during 1992-

2003 is related to their liquidity holdings and how this relationship may affect the effectiveness of the

bank lending channel. The empirical evidence provided suggests that firms holding high liquid assets

may replace bank lending by other sources of financing. Hence, higher liquidity holdings allow firms

to invest in attractive investment projects in the event of a tightening of monetary conditions. (135

words)

Key words: monetary policy transmission, liquidity, firms.

JEL Codes: E51, G21, D21.

3

1. Introduction

During the 1990s, liquidity was relatively abundant in the European Union and the European

central banks mostly developed a relaxed monetary policy. However, the demand for bank corporate

lending unexpectedly slowed down, suggesting that monetary policy actions in this area were not

completely effective. If bank credit is abundant and cheap, why do firms would rather use their

liquidity or other non-bank financing such as trade credit?1 While the main argument of the so-called

“bank-lending channel” is that a relaxed monetary policy induces banks to provide more lending, this

did not seem to happen during this period. There is empirical evidence indicating that financial

institutions with a higher proportion of liquid assets exhibit greater capacity to maintain the level of

their credit investments in the event of a hardening of monetary conditions and they do not need to

rely on other alternative sources of finance (Kashyap and Stein, 2000). Loupias et al. (2001) have also

explored these relationships for European banking over the period 1993-2000 and find that bank

lending decreases after a monetary policy tightening, although banks’ liquidity appears to affect

significantly the lending behaviour. In the case of non-financial firms, however, there is very limited

empirical evidence that these effects of liquidity may alter their financing decisions, particularly in

Europe.

This article aims to analyze the relationship between the financing behaviour of firms and their

liquidity holdings and the effects of this relationship on the effectiveness of monetary policy. The

nature of these decisions should not be trivial for policymakers, central banks, banking institutions and

non-financial firms in the context of the European financial integration and the single European

monetary policy. It should be also taken into account that European firms rely more heavily on bank

loans than their US counterparts.

This study aims to contribute to the existing literature by offering new empirical evidence on

the effects of monetary policy on firm financial structure. We employ Spain as the laboratory for this

1 Trade credit arises when a supplier allows a customer to delay payment for goods already delivered. In general,

it is associated with the purchase of intermediate goods (Cuñat, 2007).

4

exercise. The Spanish case is a particularly interesting laboratory for two main reasons: 1) the 95

percent of Spanish firms are SME and they represent more than the 65 percent of total employment.

These firms also depend critically on bank credit to undertake their investment projects and they have

a very limited access to capital markets; 2) during the sample period (1992-2003) there has been a

considerable reduction in the opportunity costs of maintaining liquid assets, as a consequence of a

substantial reduction of interest rates.

The study is organized as follows. Section 2 surveys the existing literature. The empirical

methodology is described in section 3. The main empirical goal is to analyze how the changes in the

structure of firm external finance are related to monetary policy conditions during 1992-2003. Several

structural and non-structural characteristics are considered as control variables. Section 4 shows the

main empirical results. The study ends with a summary of the main conclusions and policy

implications in section 5.

2. Firm financing and monetary policy conditions: background

According to Modigliani and Miller (1958) –under the assumption of perfect markets and

information– the market value of a firm is independent of its financial structure. Investment decisions

would then depend only on the expected rate of return. In this context, it is indifferent to firms whether

to use their own capital or to obtain external finance in order to carry out their investment projects.

Likewise, the distinction between bank debt and non-bank debt would not be relevant, as the providers

of both types of funding face the same supply conditions.

The empirical evidence shows that the perfect information model does not fit with the reality

of firms (Kashyap et al., 1994; Bernanke et al., 1996). In the presence of asymmetric information, and

given the non-perfect substitutive character of the different sources of finance, firms exhibit the

following order of preferences within the available alternatives of funding: own funds, trade credit,

capital markets financing and bank credit (Myers and Majluf, 1984; Calomiris and Hubbard, 1990).

The way in which this structure is materialised determines the composition of the balance sheet, as

5

well as the external finance premium, borne because of the cost assumed in the valuation of the

collateral offered, as well as the control carried out during the period in which a debtor position is

maintained (Stiglitz and Weiss, 1981).

Most of the previous studies have widely analysed the effects of monetary policy on the

volume of credit in the economy (Bernanke and Blinder, 1992; Bernanke and Gertler, 1995; Kashyap

and Stein, 2000). Most of them, however, have only tangentially focused on the reaction of firms to

changes in monetary policy. An exception is Kashyap et al. (1993). This study shows that while a

monetary contraction reduces bank lending, it increases commercial paper volume. This finding

suggests that monetary tightening conditions lead to an inward shift in loan supply, rather than an

inward shift in loan demand. However, it is also plausible that there is a “compositional shift” in play,

with large firms demanding more credit than small firms. Since most commercial paper is issued by

large firms, this could explain the Kashyap et al. (1993) results.

Similarly, Kashyap and Stein (2000) show that tight money should pose a special problem for

small firms, which are more likely to be bank-dependent. Therefore, contractions in policy intensify

liquidity constraints in the inventory and negatively affect the investment decisions of small firms.

While this is consistent with the lending view, there is another interpretation in what Bemanke and

Gertler (1995) called the "balance sheet channel," whereby tight monetary policy weakens the

creditworthiness of small firms, and hence reduces their ability to raise funds from any external

provider, not just banks. Given the relevance of information asymmetries in the process of credit

supply (from any lender the firm can interact with), the theory of the balance sheet channel establishes

that transformations in the structure of firms’ balance sheets –originated by the propagation of the

economic cycle– may alter their capacity to obtain and spend resources, leading to the generation of

endogenous credit cycles (Kiyotaki and Moore, 1997; Braun and Larrain, 2005). In this context, it is

firms’ financial wealth which determines their access to financing, since financial wealth acts as

collateral for the possible non-repayment of the capital contributed (Gertler, 1988). Even the strand of

6

research that focuses on the bank lending channel indirectly assumes that firm characteristics such as

financial wealth, financial debt, and some other structural financial characteristics (i.e. level of capital

and liquidity), determine the degree of access to bank credit. This is likely to occur, in particular, when

credit supply shrinks following a tightening of monetary conditions (Kashyap and Stein, 1995; Stein,

1998; Kishan and Opiela, 2000; Atanasova and Wilson, 2004). Consequently, it is expected that

smaller firms and those with a lower level of capital will be affected to a larger extent by a contractive

disturbance of monetary policy (Kashyap et al., 1993).

Firms of larger size present less severe problems of moral hazard and adverse selection

because of their a priori greater transparency. This is perceived by the markets, so that they have a

greater capacity to obtain external financing and to replace bank credit by other types of financing (at

least compared to small firms) when interest rates rise (Hubbard, 1998). Likewise, those firms with

greater own capital strength will display a greater capacity to carry out their investment projects

(Baccheta and Caminal, 2000). Consequently, it is relatively frequent to observe that the bank lending

channel operate in a similar fashion for both financial and non-financial firms when these two criteria

are explicitly considered.

A different conclusion, however, is obtained when the analysis considers the role of firm

liquidity holdings. On the one hand, there is empirical evidence that suggests that the holding of liquid

assets above a certain threshold limits the possibilities of obtaining external funding resources, since it

decreases the possibilities of assets portfolio transformation as well as the net value of the firm and the

collateral that can be offered (Morellec, 2001). On the other hand, other studies maintain that firms

with a substantial cushion of liquidity are better placed to grant and obtain finance from other firms in

the economy, especially when there have been successive falls in interest rates. The effect of liquidity

may well then explain why both the bank-lending channel and the balance-sheet channel prediction of

a reduction in lending to firms with a tightening of monetary conditions may not always occur in

practice. A recent study by Den Haan, Sumner and Yamashiro (2008) shows that real estate and

7

consumer loans sharply decreased following a monetary tightening, while commercial and industrial

loans increased. These responses, together with the responses of relevant lending rates, are hard to

reconcile with a decline in the supply of commercial and industrial bank loans during a monetary

downturn as stressed by the bank-lending channel. It is possible that liquidity and portfolio

considerations imply that the supply of commercial and industrial loans actually increases following a

monetary tightening. Even Bernanke and Gertler (1995) find that the most rapid and (in percentage

terms, by far the strongest) effect of a monetary policy shock is on residential investment, whereas

business structure investment, also a long-lived investment, does not seem to be much affected.

Trade credit between firms is a relevant variable in this context and it may also explain why

the theoretical predictions on the bank lending channel are not empirically found in many studies.

Overall, trade credit may affect the role of bank financing, in particular when firms’ liquidity holdings

are high and when firms do not get the bank loans they desire. Wilner (2000) suggests that in case of

renegotiation of debts, suppliers give more concessions to customers than banks do.2 Cuñat (2007)

shows that trade credit seems to be more prevalent when firms have low levels of liquidity; in

particular, the levels of trade credit are higher when firms experience liquidity shocks and justifies the

existence of trade credit on the basis of suppliers being able to enforce debt repayment better than

banks. This extra enforceability power comes from the link that makes both suppliers and customers

costly to substitute.

Together with firm liquidity, the banks’ attitudes toward risk in a context of low interest rates

may also explain the lack of explanatory power of the bank lending channel in certain situations.

Jiménez et al. (2007) suggest that bank risk-taking increases when interest rates are lower prior to loan

origination and that in this way monetary policy affects the composition of credit in the economy (i.e.,

the quality distribution of borrowers in the banks’ loan portfolios). They also suggest that lower

2 Frank and Maksimovic (2004) also explain the existence of trade credit as the result of suppliers having an

advantage in liquidating intermediate goods in case of default by their buyers, given that they have specialized

distribution channels to achieve this goal. These studies suggest that suppliers may then specialize in financing

buyers with low creditworthiness, for whom liquidation is more likely to occur and may act also as liquidity

providers, insuring against liquidity shocks that could endanger the survival of their customer relationships.

8

interest rates prior to loan origination result in banks granting loans with higher credit risk, but also

that banks soften their lending standards and they lend more to borrowers with a bad credit history and

with higher uncertainty. Matsuyama (2007), for example, shows that an increase of the borrowers’ net

worth (e.g. through a decrease in interest rates) reduces agency costs thus making financiers more

willing to lend to riskier borrowers (with less access to pledgeable assets). Low borrowers’ net worth,

on the other hand, may impel financiers to flee to quality (Bernanke et al., 1996). Low interest rates

may also abate adverse selection problems in the credit markets, causing banks to relax their lending

standards and increase their risk-taking (Dell'Ariccia and Marquez, 2006). In general, low interest

rates make riskless assets less attractive for financial institutions increasing their demand for riskier

assets with higher expected returns (Rajan, 2006).

3. Empirical methodology and data

3.1. Specification and definition of variables

Our empirical model relies on the standard theoretical framework of the bank lending channel

formalized by Kashyap et al. (1993). In this model, a firm selects an optimal mix of bank debt (B) and

non-bank debt (D) and seeks to minimize the financial cost. With MP denoting the stance of monetary

policy and differentiating the structure of external finance (B/D) with respect to MP yields:

( )( / ) 1

''( / )

B Nr rB D

MP f B D MP

∂ −∂=

∂ ∂ (1)

This main result of the Kashyap et al. (1993) model shows that the optimal structure of

external finance (B/D) moves inversely with the spread between bank loan rates (rB) and non-bank

debt rates (rN). A tightening of monetary conditions in this framework is then expected to reduce the

supply of bank loans relative to non-bank loans. The f in (1) shows the relationship benefit that a firm

derives from bank borrowing and it is an increasing concave function (f’>0 and f’’<0).

9

Following the framework of Kashyap et al. (1993), many extensions and empirical

applications followed where bank factors such as liquidity, loan supply and capitalization (Kashyap

and Stein, 1995; Stein, 1998; Kishan and Opiela, 2000) and firm characteristics such as financial debt,

collateral or liquidity (Benito, 2005; Guariglia and Mateut, 2006) were added to determine the degree

of access to bank credit in the event of a tightening of monetary conditions. Our empirical model

follows the standard structure of this type of empirical applications. The general equation to be

estimated is given by:

( ) ( ) ( )

( )

3 3 3 3

1 0 0 0

3

0

nt n j j t j j jn t j n t j n t j

j j j j

j t ntn t j

j

FINDEX FINDEX r ta CAP

LIQ d

α β χ δ φ

ϕ ε

−− − −= = = =

−=

∆ = + ∆ + ∆ + ∆ + ∆

+ ∆ + +

∑ ∑ ∑ ∑

∑ (2)

where FINDEX is an indicator that represents the structure of external finance, as well as the

debt maintained by firms at time t; r is the official interest rate; TA states for firm size; CAP is the

level of capitalisation; LIQ is the level of firm liquidity; and ∆, d, ϕ and ε, are the difference operator,

the vector of dummy time variables, and an error term, respectively. Finally, α is a firm fixed effect.

and d is a deterministic trend.

Three indicators were used as measures of external financial structure (FINDEX). The first of

them (BD) is defined as “Bank Debt/Non-bank Debt”. The lower this indicator is, the higher the use

of other sources of external funding other than bank loans when a firm experiences restrictions to

access to bank credit following a monetary contraction (Meltzer, 1960; Nilsen 2002). The second

indicator (D) is defined as the quotient between bank credit and total assets. The lower this indicator

is, the higher the use of capital markets instruments when the firm faces a reduction in bank credit

supply, a rise in the cost of bank credit or a higher demand of collateral (Gertler and Gilchirst., 1994;

Oliner and Rudebusch, 1995)3. Finally, the third indicator (NB) is defined as non-bank debt as a

3 In practice, the access to capital markets is usually only possible for large firms.

10

proportion of total assets (Hubbard, 1998). The lower NB is, the higher the use of equity as a response

to a monetary contraction.

The monetary policy interest rate employed -in line with the generally accepted literature

(Kashyap et al., 1994; Oliner and Rudebusch, 1996)- is the inter-bank interest rate on non-transferable

three month deposits. 4 Given the substantial dependence of Spanish firms on bank finance (Estrada

and Vallés, 1998)- the cost of bank credit is determined by the rate that intermediaries have to pay to

obtain funding- so that the opportunity cost is appropriately reflected. 5 We alternatively employed the

so-called “exogenous component of monetary policy”. This component is estimated using the reaction

function of the central bank using a VAR approach when the main identifying assumption is that

policy shocks have to be orthogonal to variables in the reaction function of the central bank. Hence,

within the system of equations in the VAR, policy shocks can be estimated as the residuals in the

linear regression of the central bank instrument on the variables in the central bank reaction function.

Various specifications of a VAR were tested using regressors such as consumer prices, the index of

industrial production and monetary policy instruments such as M2 and the 3-month T-bill rate.

With regard to the bank lending channel, those firms that exhibit a greater dependence on bank

financing, find that monetary policy decisions are more intensively reflected in their balance-sheet

structure. In order to analyse the distributive effects of this transmission mechanism it is necessary to

take into account the role played by certain specific structural characteristics of firms that may

exacerbate asymmetric information, agency costs and moral hazard problems.

Even when size does not directly determine the possibilities of access to external funding and

its cost, this variable shows a high correlation with the factors determining return risk and volatility

(Gertler and Gilchirst, 1994; Hubbard, 1998). Given the existence of asymmetric information, the size

4 We did not use the interest rate set by the Central Bank, because of the change in the public body responsible

for the application of monetary policy (from the Bank of Spain to the ECB) during our period of study. 5 The interest rate charged on firms’ credit is determined, inter alia, by monetary policy conditions, by the

market structure of the banking market, by the negotiating power of the firms, and by the existence of long-term

contractual relationships between the lender and the borrower (Berger and Udell, 2002).

11

of the firm (TA) -given by the logarithm of Total Assets- will proxy the availability of information on

the managerial capacity of firms and degree of management control (Stiglitz and Weiss, 1981). Along

with these features, TA also captures likely problems of moral hazard that arise due to the existence of

barriers to monitor investments, which are reflected in a higher external financial premium (Petersen

and Rajan, 1994).

Holding liquid assets above a certain threshold limits the possibilities of obtaining external

funding. It also diminishes the possibilities of transforming the assets portfolio, as well as the net value

of the firm and, therefore, the collateral that can be offered (Morellec, 2001). On the other hand, it

may occur that firms that operate with a substantial liquidity buffer are likely to provide funding to

other production units, in particular, in a context of low interest rates (Kim et al. 1998). The effect of

liquidity is measured by LIQ, which is defined as the ratio “cash and other highly liquid assets/total

assets”.

The financial structure of the firm is reflected by the level of capitalisation (CAP), defined as

equity relative to total assets. The existing empirical evidence has shown that firms will try to carry

out their investment projects relying as much as possible on their own resources, provided that they do

not suffer problems of decapitalisation (Baccheta and Caminal, 2000).

These structural characteristics will also help us define the main empirical hypotheses. Our

main hypothesis is that firms’ liquidity holdings will reduce the effectiveness of the bank-lending

channel. This will be our “liquidity hypothesis 1” where the effects of a monetary tightening on the

substitution of bank debt by non-bank debt are expected to decrease when liquidity increases. Size and

capitalization levels are also expected to affect the substitution of bank debt by non-bank debt. For this

reason, we will also test a “liquidity hypothesis 2” which states that liquidity holdings are expected to

reduce the effectiveness of a monetary tightening on the substitution of bank debt by non-bank debt

even when the interaction of liquidity with size and capitalization is considered.

12

3.2. Methodology and data

We employ dynamic panel data techniques to estimate equation (2). In particular, we employ

the GMM estimator of Arellano and Bond (1991) and Blundell and Bond (1998) given its capacity to

reduce the estimation bias -whatever the size of the sample- resulting from the inclusion of lags of the

dependent variable. A general specification of the model is:

( )

1

(́ ) , 1,...., ; 1,....,p

it k i t k it t i it i

k

y y L x v t q T i Nα β λ η−=

= + + + + = + =∑ (3)

where ηηηηi and λλλλt are, respectively, firm-specific and time effects; xit is a vector of explanatory

variables;, ββββ(L) is a vector associated to the differentiated variable and p is the maximum number of

lags incorporated to this model. The number of observations of each firm i is Ti,and it is higher than

the number of firms in the sample, N.

For each firm i, the reduced form of the model is:

i i i i iy W vδ ιη= + +

where δδδδ is a vector of the parameters αααα´s, ββββ´s y λλλλ´s, and Wi is a matrix that contains both the

time series of the lagged endogenous variables, the explanatory variables and the time dummies.

Finally, ιιιιi is a vector of ones of (Ti –q)x1 dimension.

Under this reduced-form specification, the GMM linear estimator of δδδδ is:

1

*´ ´ * *´ ´ˆi i N i i i i N i i

i i i i

W Z A Z W W Z A Z yδ

−

= ∑ ∑ ∑ ∑

where:

´1N i i i

i

A Z H ZN

= ∑

13

and W*

i and y*

i represent any of the likely transformations of Wi and yi (levels, first differences,

orthogonal deviations, …). Zi is a matrix of instrumental variables. Hi is s weighted matrix of firm-

specific effects.

When the number of columns in Zi equals the number of colums in W*i, AN is irrelevant and

ˆδ can be expressed as:

1

´ * ´ *ˆi i i i

i i

Z W Z yδ−

= ∑ ∑

This GMM estimator is based on a simultaneous estimation of two equations. The first one is

the regression in differences of equation (2), while the second refers to its estimation in levels. This

method provides consistent and efficient estimations, if the appropriate instruments are employed,

considering the residual correlation properties of the model (Hsiao, 1986). The use of dynamic models

requires some transformations that allow the use of lagged endogenous variables as instruments. An

efficient GMM estimation is based upon the use of a different number of instruments for each period.

In our case, these instruments are: 2 to 5 lags of the dependent variable; 1 to 5 lags of the monetary

policy variable; and 1 to 5 lags of the firm-specific characteristics, as well 1 to 5 lags of the interaction

terms between firm size (liquidity and capitalization) and the monetary policy variable.

We employ microeconomic data from the pan-European Bureau Van Dijk Amadeus database.

The sample consists of 15.617 Spanish firms for the period 1992-2003, resulting in a panel of 136.247

observations.6,7,8

The data correspond to the consolidated accounting statements, since we aim to

reflect possible transfers of assets or liquidity between firms that belong to a single business holding.

6 We did not employ macroeconomic data since these may produce biased results on the operativity of the

transmission channels of monetary policy due to: 1) simultaneity problems; 2) frictions in the capital markets;

and 3) heterogeneity among the firms in the sample (Chirinko et al., 1999). 7 This database contains information on, inter alia, balance sheet structure, profit and loss account, number of

employees, legal nature and industry classification. However, there was a significant lack of data on the variables

showing firm age and credit rating so that we were not able to incorporate them into the analysis. 8 The periodicity of the data is annual, so we must bear in mind the potential limitations on the analysis due to

the impossibility of reflecting: 1) the immediate impact of variations in interest rates on the composition of

external finance; and 2) the bias in the composition of the sample, as a consequence of the predominance of large

firms.

14

Finally, the inter-bank interest rate on three month transferable deposits was taken from the Statistical

Bulletin of the Bank of Spain. The descriptive statistics and the correlation matrix of the variables are

shown in Tables 1 and 2, respectively.

In order to test the hypotheses derived from the theory of the bank lending channel, equation

(2) is estimated to analyze the effects of interest rates on the various debt ratios. Additionally, each of

the equations incorporates cross dummies (interaction terms) between interest rates (r), size (ta)

liquidity (LIQ), and capitalization (CAP) to analyze whether monetary policy measures generate

distributive effects. As for size, firms are considered as “large” if they belong to the last quartile, while

the “small” ones are those in the first quartile.9 The same criterion was applied to classify firms in

terms of capitalisation and liquidity.

4. Empirical evidence.

The regression analysis was carried out considering the three specifications of FINDEX as

dependent variable. We estimate the long-term coefficients (η), which are given by the sum of the

short-term coefficients of each of the independent variables, divided by one minus the sum of the short

term coefficients of the dependent variable (Chatelain et al., 2004): 10

∑

∑

=

=

β−

Φ

=η3

1

3

0

1i

i

i

i

(2)

where Φ represents χ, δ, φ, ϕ, and γ, respectively. 11

9 The distribution of firms among the different categories has been undertaken for each individual period, in

order to reflect the dynamic nature of the data. Consequently, each firm may appear in different classifications

for each year, so the number of firms in the various categories need not remain constant throughout the period

considered. 10

These coefficients represent the long-term percentage change in the indicator of the composition of external

finance or debt, in response to a permanent variation of 1 % of any explanatory variables (i.e. interest rates, size,

degree of capitalisation, liquidity, or bank gearing of the firm). 11 We have not included the short-term coefficients mainly because these coefficients have lower economic

explanatory power and they are largely affected by strictly conjectural factors. The exclusion of these results also

simplifies the empirical evidence of the study.

15

The values of the Sargan test confirm the validity of the instruments employed. 12

On other

hand, the values corresponding to the AR1 and AR2 tests indicate that there is no second-order

autocorrelation. The Huber-White procedure was also employed, in which the standard errors are

calculated on the basis of the quasi-verisimilitude function.

4.1. The impact of monetary policy measures.

Table 3 summarizes the main results of the impact of monetary policy actions. A unit increase

in the inter-bank interest rate results in a reduction of .211 in BD (Bank Debt/Non-bank Debt). This

finding seems to confirm that firms turn to non-bank debt (ie. trade credit) to obtain the necessary

resources for their business in the presence of imperfections in the bank credit market. This evidence

is in line with the results obtained for the USA by Kashyap et al. (1994) and Oliner and Rudebusch

(1996). As for the effects of size, capitalization and liquidity, a monetary tightening permits a larger

substitution of bank debt by non-bank debt at larger firms as well as at firms showing higher liquidity

and capitalization levels. Interestingly, the impacts of size, capitalization and liquidity on firm

financing decisions seem to be significantly higher than that of interest rates. The coefficient of size is

-1.768. As for capitalization, the coefficient is -1.456, suggesting that the availability of sufficient

internal funds to undertake investment projects reduces the use of external finance (Baccheta and

Caminal, 2000). Similarly, a 1% increase in the relative weight of liquid assets on the balance sheet

reduces the ratio BD by -1.312, a result that seems to support our “liquidity hypothesis 1”.

The comparison of the coefficients of the interaction terms (using mean differences test, not

shown) between the explanatory variables in Table 3 suggests that differences in size may be less

important once liquidity and capitalization are controlled for. Similarly, this also happens to

capitalization once size and liquidity are controlled for. However the impact of liquidity on the

substitution of bank debt by non-bank debt is statistically significantly higher for smaller firms as well

as for those showing lower levels of capitalization (not shown). All in all, this evidence suggests that

12

The values obtained by incorporating a smaller number of instruments were much poorer, as a consequence of

the complex structure of the model (Chatelain et al., 2004).

16

there is a limited role for the bank lending channel of monetary policy when liquidity holdings are

high because firms may replace bank lending by other sources of funding even when size and liquidity

are considered, which supports our “liquidity hypothesis 2”.

The findings are similar when we employ D (Bank credit/total assets) as dependent variable. A

unit increase in interest rates leads to a reduction of .243 in D in the long-term. Larger firms and those

with higher capitalization and liquidity seem to make a more intensive use of non-bank finance, due to

its lower cost. Finally, a rise in interest rates does not seem to generate a substantial and significant

reduction in NB (non-bank deb/total assets). The long-run coefficient of r in this case is -.041. This

result reinforces the hypothesis that non-bank debt is less sensible to changes in interest rates, thereby

driving firms to increase their proportion of non-bank debt relative to bank loans when monetary

conditions are tougher.

4.2. Robustness checks

In order to test the robustness of the empirical results we employed alternative variables as

well as various aspects that may determine firms’ financial behaviour as a reaction to changes in

monetary policy. 13 First of all, we replaced the three month inter-bank interest rate by the deviation of

the interest rate from the rate estimated by the reaction function of the Central Bank (obtained

applying a VAR methodology). The results confirm that firms tend to alter the composition of their

debt when monetary conditions tighten. The magnitude of this effect is slightly lower (1 %), probably

as a consequence of the underestimation of the exogenous component of the interest rate which is

frequently observed when applying this methodology (Bernanke and Mihov, 1998). Secondly, we

incorporate the cost of the debt in order to capture the possible influence of the relationship of the firm

with the supplier of funding (i.e. how rates are set for the firm and how the rates are affected in a

monetary contraction). The results suggested that the relationship was not statistically significant. We

also include the ratio “profits before taxes/interest paid”, in order to quantify the restrictions set by

firms’ profitability on firms’ financial decisions. The results suggested that this factor did not have a

13

The results are available on request to the authors.

17

significant impact. A similar conclusion was obtained when this measure is replaced by cash flow

(defined as cash receipts minus cash payments). In both cases, the long-term coefficient was not

statistically significant. The inclusion of ROA (Return on Assets) or ROE (Return on Equity) does not

show that profitability had significantly influenced the composition of firms’ debt. Only ROA was

statistically significant and positive when the regression analysis was carried out for the total sample,

and the dependent variable was BD. The logarithm of GDP and the logarithm of firms’ sales were also

introduced in order to capture a likely impact of the economic cycle. The effects of these variables

were not found to be significant at 5% level.

Equation (2) was estimated separately for seven different industries (Agriculture, Hunting and

Fishery; Extraction Industry, Energy & Water; Manufacturing Industry, Retail trade, Repairs,

Domestic articles, Hotel, Restaurants, Transport and Communications; Construction; Other

marketable services). Since we found no significant statistical differences across industries, we have

not included these results for the sake of simplicity. Additionally, the linear relationship shown in

equation (2) was modified by considering some quadratic terms on r, ta, CAP and LIQ. However,

none of these quadratic terms was found to be statistically significant (not shown) and, therefore, non-

linear relationships do not seem to alter the results in this case.

5. Conclusions.

During the 1990s, liquidity was relatively abundant in the European Union and European

central banks followed relaxed monetary policies. However, the demand for corporate bank lending

slowed down, suggesting that monetary policy was not totally effective in this area. In this context,

firms may have decided to hold liquid assets to protect themselves against future unfavourable

scenarios such as growing interest rates, lower earnings and higher restrictions and costs of accessing

capital markets. This article analyses the empirical relationship between liquidity holdings and firm

financial structure in order to assess how monetary policy affects firm financing in this context. This

relationship is tested on a sample of Spanish firms during 1992-2003. Using dynamic panel data

18

techniques, the empirical results show that when interest rates increase, firms reduce their dependence

on bank lending and maintain a higher level of liquidity. These results are in line with previous studies

suggesting that there is a limited role for the bank lending channel of monetary policy when liquidity

holdings are high because firms may replace bank lending by other sources of funding. We also find

that non-bank debt is less sensible to changes in interest rates, thereby driving firms to increase their

proportion of non-bank debt relative to bank loans when monetary conditions are tougher. As noted by

Kashyap et al. (1993), this finding may imply that there is a “compositional shift” in play, with large

firms demanding more credit than small firms. Although the empirical model does not directly test the

effects of trade credit on monetary policy, our results are indirectly in line with recent research on the

role of trade credit as a substitute for bank loans and how this substitution may affect the relevance of

the bank lending channel (Frank and Maksimovic, 2004; Cuñat, 2007).

Further research is needed on the impact of the monetary policy on trade credit as well as on the

effects of trade credit financing on firms investment. Additionally, future research should also

consider the effects of bank ownership structure and the legal and institutional characteristics of the

financial markets in which the firm operates to provide with a more accurate picture of the effects of

liquidity, size or capitalization on the effectiveness of the bank lending channel. More work is also

needed in order to test how a more relaxed (or tougher) monetary policy may respond to additional

factors from the supply side and the behavior of banks, as we can experience from financial crises

nowadays.

Considering all these findings and potential research avenues, regulators should seriously

consider the variety of mechanisms that firms are employing to obtain financing in the presence of

diverse monetary conditions. In particular, the European Central Bank should consider situations in

which expansive monetary policies will be hardly effective to promote firm investments and the so-

called bank lending channel of monetary policy may be less effective than expected. Regulators and

monetary authorities will also need to have a closer look at non-bank financial intermediaries as well

as firms acting as trade creditors since they may diminish the role of the bank lending channel and new

19

paradigms are needed to infer the economic impact of monetary conditions. All these financing

mechanisms also pose some implications for risk monitoring since most of the regulatory pressure is

placed on banks while other non-bank financing suppliers are also developing rapidly in many

markets.

20

References

Arellano, M., and Bond, S. (1991). “Some test specifications for panel data: Monte Carlo evidence and

an application to employment equations”, Review of Economic Studies, 58: 277-297.

Atanasova, C. and Wilson, N. (2004). “Disequilibrium in the UK corporate loan market”, Journal of

Banking and Finance, 28: 595–614.

Baccheta, P., and Caminal, R. (2000). “Do capital market imperfections exacerbate output

fluctuations?, European Economic Review, 44: 449-468.

Benito, A. (2005). “Financial pressure, monetary policy effects and inventories: Firm-level evidence

from a market-based and a bank-based financial system”, Economica 72: 201–224.

Berger, A. and Udell, G. (2002). “Small business credit availability and relationship lending: the

importance of bank organization structure”, Economic Journal, 112: 32-53.

Bernanke, B.S., and Gertler, M. (1995), "Inside the Black Box: The Credit Channel of Monetary

Policy Transmission", Journal of Economic Perspectives 9, 27-48.

Bernanke, B., Gertler M. and Gilchrist, S. (1996). “The financial accelerator and the flight to quality”,

Review of Economics and Statistics, 78: 1-15.

Bernanke, B.S., and Blinder, A.S: (1992), "The Federal Funds Rate and the Channels of Monetary

Transmission", American Economic Review 82, 901-921.

Bernanke, B. and Mihov, I. (1998). Measuring monetary policy, Quarterly Journal of Economics,

113: 869-902.

Blundel, R. and Bond, S. (1998). “Initial conditions and moments restrictions in dynamic panel data

models”, Journal of Econometrics, 87: 115-143.

Braun, M. and Larrain, B. (2005). “Finance and the business cycle: International, inter-industry

evidence”, Journal of Finance, 60: 1097–1128.

Calomiris, C. and Hubbard, R. (1990). “Firm heterogeneity internal finance and credit rationing”,

Economic Journal, 100: 90-104.

Chatelain, J., Hernando, I., Von Kalckreuth, U. and Vermeulen, P. (2004), “Firm investment and

monetary transmission in the euro area”, in: (I. Angeloni, et al., eds.), Monetary policy transmission

in the Euro area, Cambridge University Press, Cambridge.

Chirinko, R., Fazzari, S. and Meyer, A. (1999). “How responsive is business capital formation to its

user cost?”, Journal of Public Economics, 74: 53-80.

Cuñat, V. (2007), “Trade Credit: Suppliers as Debt Collectors and Insurance Providers, Review of

Financial Studies” 20: 491-527.

Den Haan, W.J., S. Sumner, and Yamashiro, G. (2008), "Bank Loan Portfolios and the Monetary

Transmission Mechanism," Journal of Monetary Economics, 55, forthcoming.

Dell'Ariccia, G., and Marquez, R. (2006), "Lending Booms and Lending Standards", Journal of

Finance 61, 2511-2546.

21

Diamond, D. (1984). “Financial intermediation and delegated monitoring”, Review of Economic

Studies, 54: 393-414.

Diamond, D. (1991). “Monitoring and reputation: the choice between bank loans and directly placed

debt”, Journal of Political Economy, 99: 689-721.

Estrada, A. and Vallés, J. (1998). “Investment and financial system structure in Spanish manufacturing

firms”, Investigaciones Económicas, 22: 337-359.

Frank, M., and V. Maksimovic (2004), ‘‘Trade Credit, Collateral, and Adverse Selection,’’ Mimeo

University of Maryland.

Gertler, M. (1988). “Financial structure and aggregate economic activity: an overview”, Journal of

Money, Credit, and Banking, 20: 559-588.

Gertler, M. and Gilchrist, S. (1994). Monetary policy business cycles and the behaviour of small

manufacturing firms, Quarterly Journal of Economics, 109: 309-340.

Guariglia, A. and S. Mateut (2006): “Credit channel, trade credit channel, and inventory investment:

Evidence from a panel of UK firms”, Journal of Banking and Finance, 30: 2835-2856.

Hubbard, G. (1998). “Capital market imperfections and investment”, Journal of Economic Literature,

36: 193-225.

Hsiao, C. (1986). The analysis of panel data, Cambridge University Press, Cambridge.

Jiménez, G., Saurina, J., Ongena, S. and Peydró, J.L. (2007), “Hazardous times for monetary policy:

what do twenty-three million bank loans say about the effects of monetary policy on credit risk?,

CentER working papers, 2007-75. Tilburg, Netherlands

Kashyap, A., Lamont, O. and Stein, J. (1994). “Credit conditions and the cyclical behaviour of

inventories”, Quarterly Journal of Economics, 109: 565-592.

Kahsyap, A. and Stein, J. (1995). “The impact of monetary policy on bank balance sheet”, Carnegie

Rochester Conference Series on Public Economics, 42: 151-192.

Kashyap, A., Stein, J. and Wilcox, D.W. (1993). “Monetary policy and credit conditions: evidence

from the composition of external finance”, American Economic Review, 83: 78-98.

Kashyap, A. and Stein, J. (2000), “What do a million observations on banks say about the transmission

of monetary policy”, American Economic Review, 90: 407-428.

Kim, C., Mauer, D. and Sherman, E. (1998), “The determinants of corporate liquidity: theory and

evidence”, Journal of Financial and Quantitative Analysis, 33: 335-359.

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

Money, Credit, and Banking, 32: 121-141.

Kiyotaki, N. and Moore, J. (1997), “Credit cycles”, Journal of Political Economy, 105: 211-248.

Loupias, C., Savignac, F. and Sevestre, P.. (2001).”Monetary policy and bank lending in France: are

there asymmetries?”, European Central Bank Working Paper Series, 101, 2001.

Meltzer, A. (1960). “Mercantile credit, monetary policy, and the size of the firms”, Review of

Economics and Statistics, 42: 429-437.

22

Modigliani, F. and Miller, M. (1958). “The cost of capital, consumption finance, and the theory of

investment”, American Economic Review, 48: 261-297.

Morellec, E. (2001). “Asset liquidity, capacity choice and the pricing of corporate securities”, Journal

of Financial Economics, 61: 173-206.

Myers, S. and Majluf, N. (1984). “Corporate financing and investment decisions when firms have

information that investors do not have”, Journal of Financial Economics, 13: 187-221.

Myers, S. and Rajan, R. (1998). “The paradox of liquidity”, Quarterly Journal of Economics, 113:

733-771.

Nilsen, J. (2002). “Trade credit and the bank lending channel”, Journal of Money, Credit and Banking

34: 226-253.

Oliner, S. and Rudebusch, G. (1995). “Is there a bank lending channel of monetary policy?”, Federal

Reserve Bank of San Francisco Economic Review, 2: 3-20.

Oliner, S. and Rudebusch, G. (1996). “Monetary policy and credit conditions: evidence from the

composition of external finance: comment”, American Economic Review, 86: 300-309.

Petersen, M. and Rajan, R. (1994). “The benefit of lending relationships: evidence from small business

data”, Journal of Finance, 49: 421-460.

Rajan, G.R. (2006), "Has Finance Made the World Riskier?,"European Financial Management” 12,

499-533.

Sharpe, S. (1991). “Asymmetric information, bank lending and implicit contracts: a stylized model of

customer relationship”, Journal of Finance, 45: 1069-1087.

Stein, J. (1998). “An adverse selection model of bank asset and liability management with

implications for the transmission of monetary policy”, Rand Journal of Economics, 29: 466- 486.

Stiglitz, J. and Weiss, A. (1981). “Credit rationing in market with imperfect information”, American

Economic Review, 71: 393-410.

Wilner, B. S. (2000), ‘‘The Exploitation of Relationships in Financial Distress: The Case of Trade

Credit,’’ Journal of Finance, 55, 153-178.

23

Table nº1: Summary Statistics

Mean Std. Dev Maximun Minimun

BD .82 .60 1.20 .15

B(%) 28.80 25.30 95.5 3.52

NB (%) 35.18 27.90 90.15 5.13

ta (thousand euros) 113,487 957,398 47,936,395 77

CAP(%) 38.5 14.5 79.90 8.3

LIQ(%) 11.1 16.0 22.0 3.0

r(%) 5.60 3.36 14.51 4.03

Table nº2: Correlation matrix

BD B NB ta LIQ CAP r

BD 1 .72 -.80 -.62 -.45 -.61 -.58

B .72 1 -.69 -.51 -.35 -.41 -.47

NB -.80 -.69 1 -.57 -.63 -.47 -.41

ta -.62 -.51 -.57 1 .42 .54 -.39

LIQ -.45 -.35 -.63 .42 1 .40 -.19

CAP -.61 -.41 -.47 .54 .40 1 -.41

r -.58 -.47 -.41 -.39 -.19 -.41 1

24

Table nº3: Regression Results. Long-term coefficients.

Dependent variable: BD B NB

r -.211***

(-2.78)

-.243***

(-3.32)

-.041**

(-2.33)

Ta -1.768*** (-2.99)

-1.457*** (-3.46)

-.678** (-2.09)

CAP -1.456***

(-3.11)

-1.567***

(-3.21)

-.567**

(-2.25)

LIQ -1.312** (-2.39)

-1.630*** (-2.75)

-.595** (-2.00)

r effects for:

Large Firms -.367***

(-3.07)

-.302**

(-2.09)

-.031**

(-2.21)

Small Firms -.187**

(-2.34)

-.211**

(-2.25)

-.045*

(-1.94)

High Liquidity Firms -.431**

(-1.96)

-.245*

(-1.86)

-.032*

(-1.69)

Low Liquidity Firms -.243***

(-3.11)

-.119**

(-2.51)

.045*

(-1.67)

High Capitalization Firms

-.345*** (-3.23)

-.119** (-2.03)

-.032* (-1.57)

Low Capitalization Firms -.234***

(-2.78)

-.097***

(-2.23)

-.045**

(-1.99)

ta effects for:

High Liquidity Firms -.456*

(-1.66)

-.345*

(-1.92)

-.098*

(-1.79)

Low Liquidity Firms -.476**

(-2.22)

-.387**

(-2.50)

.099*

(-1.67)

High Capitalization Firms

-1.765***

(-4.16)

-.875**

(-2.11)

-.045**

(-2.17)

Low Capitalization Firms -1.711***

(-2.79)

-.999***

(-2.21)

-.065**

(-1.96)

CAP effects for:

Large Firms -1.567***

(-2.89)

-1.623**

(-2.19)

-.259**

(-2.31)

Small Firms -1.455**

(-1.96)

-1.645**

(-2.35)

-.316**

(-2.11)

High Liquidity Firms -1.643***

(-3.06)

-1.230*

(-1.72)

-.300*

(-1.87)

Low Liquidity Firms -1.771***

(-3.32)

-1.311*

(-1.90)

.212*

(-1.67)

LIQ effects for:

Large Firms -1.135***

(-2.90)

-1.044**

(-2.19)

-.553**

(-2.41)

Small Firms -1.547***

(-2.99)

-1.111**

(-2.41)

-.321**

(-1.99)

High Capitalization Firms -1.259**

(-1.96)

-1.133**

(-2.22)

-.209**

(-1.99)

Low Capitalization Firms -1.778**

(-2.22)

-1.121*

(-1.90)

.251**

(-2.37)

Residual autocorrelation test

AR1 .000 .000 .000

AR2 .243 .196 .0199

Sargan Test (2-step) .177 .201 .189

Number of firms 15,617

Number of observations 153,804

Period 1992-2003

Notes: ***/**/* denotes significance at 1, 5 and 10 % levels. t-statistics reported in parenthesis.

Estimation by GMM-system estimator using the robust two-step method. The Sargan test is a test of over-identifying

restrictions (p-value reported), distributed as chi-squared under the null of instruments validity. ARj is a test of jth-order serial

correlation in the first-differenced residuals. These are both distributed as a standard normal under the null hypothesis.