Multistrategic Behaviour of Croatian Banks Sanja Jakovljevi Introduction

Multistrategic Behaviour of Croatian Banks Sanja Jakovljević

Introduction p To what extent do branching decisions and interest rate choices matter for profit maximization in the deposit and loan market? – the case of commercial banks operating in Croatia during the period between 2004 and 2010 p Two potential novelties: The inclusion of regional deposit and loan volumes in the estimation process, as well as banks’ branching choices (the Croatian case) n Assessment of the impact of the increase in quality of transportation infrastructure n

The banking industry from structure to strategies p Structural (SCP, efficiency-structure) and non-structural (NEIO) approaches Monopolistically competitive industry n Largely confirmed for the Croatian banking sector as well n p Salop (1979) as the basis for spatial models and the multimarket nature of banks’ performance p Further steps: multistrategic and multi-output considerations p Barros (1999); Kim, Lozano-Vivas et al. (2007)

Spatial competition model Application of Salop's model to bank branch distribution Source: Adjusted from Salop (1979) p Branch b and neighboring branches b-1 and b+1, the distance between them set as db, b-1 and db, b+1 p Assumption 1: interest rates on deposits and loans uniform across regions p Assumption 2: symmetrical distribution of branches

Spatial competition model (2) p Based on the indifferent client’s location and probabilities of branches belonging to banks i or j, demand for deposits and loans faced by bank i on the regional market k is and r. . interest rate; n. . number of branches; . . volume of deposits; . . volume of loans; . . transportation costs

Spatial competition model (3) p Profit maximization where p , and First-order conditions, with respect to deposit interest rate, loan interest rate and number of branches

Spatial competition model (4) p Stage 1. Estimation of marginal costs, using the transcendental logarithmic (translog) function form and the related cost share equations, under conditions of symmetry and homogeneity of degree one in input prices

Spatial competition model (5) p Stage 2. Estimation of transportation costs, but with the use of regional volumes of deposits and loans n Estimation of a system of three first-order conditions of profit maximization simplified into the estimation of one first-order condition

Spatial competition model (6) p Elasticity coefficients DEPOSITS LOANS Bank i's interest rate on deposits Bank i's interest rate on loans Bank j's interest rate on deposits Bank j's interest rate on loans Bank i's branch network Bank j's branch network

Data and estimation methods p CNB Monthly report on interest rates on new and renewed loans and deposits; monetary policy rates; money market rate n Quarterly profit-and-loss account, balance sheet, and report on regional volumes of deposits and loans n Semi-annual report on the number of branches in each region n p CBS n p Size of regions, percentage of urban population CATMC n Size of the highway network

Data and estimation methods (2) p Stage 1. FIML method (invariant to the choice of the cost share equation to be dismissed) L. . personnel expenses; K. . expenses on asset depreciation and other administrative expenses; PL. . ratio of L to number of employees; PK. . ratio of K to the value of fixed assets

Data and estimation methods (3) p Stage 2. GLS method with cross-section weights area. . surface of region in km 2 ; urban. . percentage of urban population; hway. . length of highway network in km p The model (panel data on 32 banks between Q 1 2004 and Q 4 2010) was estimated for the whole observed period, as well as for two sub-periods: pre-crisis period (Q 1 2004 – Q 2 2008) n crisis/post-crisis period (Q 3 2008 – Q 4 2010) n

Results First and second estimation stage Note: parameters marked with * are multiplied by 10 -3, and parameters marked with ** are multiplied by 10 -5.

Results Estimated demand elasticity coefficients

Results (2) p Demand for deposits and loans is more sensitive to the banks' choice of interest rates than the change of the branch network size p Demand for loans is less sensitive to price changes than demand for deposits (availability of substitutes; refinancing purpose; nature of contracts), as is also the case for branching decisions p The crisis/post-crisis period brought about a decrease in the price sensitivity of depositors and its increase in the case of borrowers p Branching choices matter more for multimarket banks

Strategic variables of Croatian banks Source: CNB

Conclusions p Banks in Croatia seem to be more focused on interest rate choices p A few multimarket banks rely more heavily on branching choices than other banks n p Possible regulatory implications Influence of infrastructural improvements on transportation costs

Thank you for your attention!