Abstract
This chapter offers a synthesis of the characteristics of the demand and supply of credit at the regional level in Italy. The various analyses are conducted using data from the Bank of Italy, ISTAT (Italian National Institute of Statistics) and Prometeia, and cover the 2010–2014 period. In particular, the loan supply for northern regions seems more proportionate and readily responsive to household and firm needs. The interdependence between the dynamics of price formation and credit quality, captured by a vector autoregression for panel regressions, is also significant, with an intensity that varies according to loan maturity and the evolution of economic and financial variables. This result highlights an amplified, negative relationship between interest spreads and credit quality.
JEL classification: G20, G21, G28
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
It is useful to consider, for example, the quality of the credit brokerage that could certainly also be included in the more general environmental effect, as well as other infrastructure useful to realize suitable conditions to facilitate investment and profitability of production activities.
- 2.
Starting from the matrix of the added values, based on data from regional economies firm Prometeia, which shows the value added for each partition area (line) and production sector (column) to capture the geographical and sectorial specialization, the LB specialization index (Lo Cascio et al. 2012) has been calculated as:
$$ L{B}_{i,j}=\frac{q(x)-q(a)}{\left[1-q(a)\right]q(x)+\left[1-q(x)\right]q(a)}, $$where
$$ q(x)=\frac{x_{i,j}}{{\displaystyle {\sum}_i}{x}_{i,j}} $$$$ q(a)=\frac{{\displaystyle {\sum}_j}{x}_{i,j}}{{\displaystyle {\sum}_i}{x}_{i,j}} $$where x i, j is the value (loans provided in 2012) for the jth variable (sector) for the region i forq(x) = q(a) , LB i , j = 0; for q(x) < q(a) , LB i , j < 0; and for q(x) > q(a) , LB i , j > 0, 1 ≥ LB i , j ≥ − 1.
If LB presents positive values close to 1, this indicates a specialization for region i in sector j. Conversely, negative values close to –1 indicate a despecialization. The result of calculating the difference between the indicator at time t and at time t − 1 defines the dynamics of specialization.
- 3.
As calculated in the work of Jahn et al. (2013).
- 4.
For example, the excessive concentration in the credit line class is typically attributed to consumer credit (e.g., under 25,000.00 EUR) and exposes its banking sector to trends in household consumption.
- 5.
- 6.
Four clusters were chosen to better explain the results and to isolate the regions with extreme values.
- 7.
See Table A.9.1 in Appendix A for more detailed information.
- 8.
These regions present values below average for the year 2014 with regards to the indicator of loan concentration by borrower size.
- 9.
The first eight regions with the highest levels of the degree of loan concentration by size, except Basilicata, are included in the second cluster, which represents almost all of the southern regions.
- 10.
See Fig. A.9.5 in Appendix A for households.
- 11.
- 12.
If the trend of the degree of loan concentration by size of borrowers were used as a proxy of the bargaining power of banks, some useful information could be compared with estimates made on pricing policies adopted by banks at various stages of the relationship. For more on this question, see the work of Parigi (2000).
- 13.
The decision to construct a regression-pull type with fixed effects relates to the need to assign the peculiarities of origin to a regional banking system and, therefore, a different intercept. Moreover, the specificity of each appears amply justified in light of the results from previous calculations, and particularly the analysis of clusters.
- 14.
- 15.
- 16.
Regarding the analysis of correlation statistic, see Appendix B, as it is not possible to obtain appropriate information to improve the above proposed interpretative framework.
References
Beretta, E. (2004). I divari regionali tra i tassi bancari in Italia. Banca Impresa Società, 3/2004, 565–584.
Bianchi, T. (2015). L’economia d’azienda, la disciplina della mia vita accademica e didattica. In L. Guatri (Ed.), Economia Aziendale. Come era e come è. Milano: Egea.
Gallo, M. and Vannoni, V. (2015) ‘Le determinanti dei tassi di interesse bancari nelle regioni italiane’, Journal of Regional Science, vol. 14, pag. 67-84.
Jahn, N., Memmel, C., & Pfingsten, A. (2013). Banks’ concentration versus diversification in the loan portfolio: New evidence from Germany (Discussion paper, Deutsche Bundesbank No 53/2013).
Johnson, R. A., & Wichern, D. W. (2007). Applied multivariate statistical analysis. New York: Pearson.
Lo Cascio, M., Bagarani, M., & Zampino, S. (2012). Economic space trajectory through different regional growth models. In M. Bagarani (Ed.), Il governo delle Regioni e lo sviluppo economico. Limiti e rischi del processo di decentramento comunitari. Alessandria: Edizioni dell’Orso.
Mattesini, F., & Messori, M. (2004). L’evoluzione del sistema bancario meridionale: problemi aperti e possibili soluzioni. Bologna: Il Mulino.
Parigi, B. M. (2000). La concorrenza nel settore bancario: una rassegna della letteratura. In M. Polo (Ed.), Industria bancaria e concorrenza. Bologna: Il Mulino.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix A
Appendix A
Index of production specialization in Construction, 2010–2014 (the x-axis projects the specialization index in 2010, the y-axis reports the same index in 2014)
Index of production specialization in Services, 2010–2014 (the x-axis projects the specialization index in 2010, the y-axis reports the same index in 2014)
Index of loan specialization in Construction, 2010–2014 (the x-axis projects the specialization index in 2010, the y-axis reports the same index in 2014)
Index of production specialization in Services, 2010–2014 (the x-axis projects the specialization index in 2010, the y-axis reports the same index in 2014)
Credit quality for households, 2010–2014 (the x-axis projects the ratio between non-performing loans and total loans in 2010, the y-axis reports the same ratio in 2014)
Interest rate spreads with maturity between 1 and 5 years for households, 2010–2014 (the x-axis reports the value of the interest rate spread in 2010, the y-axis reports the same ratio in 2014)
Interest rate spreads with maturity up to 1 year for households, 2010–2014 (the x-axis reports the value of the interest rate spread in 2010, the y-axis reports the same ratio in 2014)
Interest rate spreads with maturity over 5 years for households, 2010–2014 (the x-axis reports the value of the interest rate spread in 2010, the y-axis reports the same ratio in 2014)
Interest rate spreads with maturity between 1 and 5 years for family businesses, 2010–2014 (the x-axis reports the value of the interest rate spread in 2010, the y-axis reports the same ratio in 2014)
Interest rate spreads with maturity up to 1 year for family businesses, 2010–2014 (the x-axis reports the value of the interest rate spread in 2010, the y-axis reports the same ratio in 2014)
Interest rate spreads with maturity over 5 years for family businesses, 2010–2014 (the x-axis reports the value of the interest rate spread in 2010, the y-axis reports the same ratio in 2014)
1.1 Appendix B: Regression model
The relationship between loan quality and the difference between active and passive interest rates is valid in either direction. On the one hand, the increase in the spread practiced allows (owing to the mechanism of adverse selection) financing only riskier investments that significantly impact the quality of credit; on the other hand, a deterioration in credit quality could increase lending rates compared to passive rates in recovering losses because of the deteriorating quality of loans. It becomes useful at this point to investigate the statistical significance and intensity of the relationships in the context of Italian regions.
A first relationship between credit quality and spreads can be specified as (Spec. A), in line with the above definition:
where q t , calculated as the ratio of non-performing loans and total loans, is the quality of loans issued at time t; s <1y , t − 1 is the spread between lending rates on loans, with a maturity up to 1 year and deposit rates at time t − 1; s 1_5y , t − 1 is the spread between lending rates on loans with a maturity between 1 and 5 years and deposit rates at time t − 1; and s <5y , t − 1 is the spread between lending rates on loans with a maturity over 5 years and borrowing rates at time t − 1. The regression coefficients and relative significance indicate the sensitivity of the credit quality to variations in spreads for different maturities and, according to the aforementioned reasoning, this should have a positive coefficient, or an increase in the spread at time t − 1 should correspond to an increase in credit risk.
The relationship can be reversed, as anticipated, by placing the quality of credit as the independent variable, delayed for a period with respect to the dependent one; the latter is considered according to the maturity of the spread (Spec. B):
The expected relationship between q t − 1 and the spread is positive; namely, a decrease in quality credit (q t − 1 increases) should determine an increase in the spread, increasing net interest, which offsets the losses associated with increased riskiness.
The dependent variables expressed in Eq. (9.1) and Eq. (9.2) may also be affected by economic situation. In fact, with regard to Eq. (9.1), in favourable economic conditions credit quality should improve, while conversely, in unfavourable economic conditions the credit quality should worsen. With regard to Eq. (9.2), the relationship between economic conditions and spreads for the banks is not known in advance, as on the one hand, banks are inclined to reduce the spread due to improvement in credit quality, and on the other hand, riskier projects in positive economic conditions may prove more profitable. It is useful in light of the aforementioned, and with these specifications, to insert a variable that summarizes the economic situation; for this reason, the variable Δy t − 1 is constructed as the logarithmic difference in GDP.
The level of spread, in addition to the economic situation captured by the GDP, may be influenced by conditions that characterize the financial market, such as the yield on government bonds, which is an opportunity to invest in a business that would be less risky for a bank than lending. If the government bond yield is high, the bank should increase the spread to ensure the same risk-adjusted return. However, by equally adopting investment logic in the financial market for the medium to long term, an increase in government bond returns could lead the bank to reduce the active interest rate to finance less risky assets, offsetting the loss of interest income with higher yields on government bonds. Specifications (9.1) and (9.2) receive another variable for these reasons, which summarizes the trend rate for alternative loan investments: the rate of yield for Italian state bonds in 10 years. The following are the results of processing structures for Specs. A and B, sorted by type of borrower:
Rights and permissions
Copyright information
© 2017 The Author(s)
About this chapter
Cite this chapter
Malavasi, R., Aliano, M. (2017). Credit Supply and Bank Interest Rates in the Italian Regions. In: Rossi, S. (eds) Access to Bank Credit and SME Financing. Palgrave Macmillan Studies in Banking and Financial Institutions. Palgrave Macmillan, Cham. https://doi.org/10.1007/978-3-319-41363-1_9
Download citation
DOI: https://doi.org/10.1007/978-3-319-41363-1_9
Published:
Publisher Name: Palgrave Macmillan, Cham
Print ISBN: 978-3-319-41362-4
Online ISBN: 978-3-319-41363-1
eBook Packages: Economics and FinanceEconomics and Finance (R0)