The Impact of Organizational Structure

1
The Impact ofOrganizational Structure Lending
Technology on Banking Competition

• Hans Degryse
• CentER - Tilburg University, TILEC, K.U. Leuven
  CESIfo
• TILEC-AFM Chair on Financial Market Regulation
• Luc Laeven
• IMF CEPR
• Steven Ongena
• CentER - Tilburg University, TILEC CEPR
• Con ference on The Changing Geography of
  Banking, Ancona, September 2006

2
Motivation Bank Organization Spatial Pricing

• Banks internal organization matters for lending
• Stein (JF 2002) Takats (ECB 2004) Liberti
  (2004) Liberti Mian (2005), Berger et al. (JFE
  2005)
• Opaque (small) firms borrow from close banks
• Petersen Rajan (JF 2002) Saunders Allen
  (2002)
• Large, centralized banks lend to distant, large
  firms using hard information
• Berger, Miller, Petersen, Rajan Stein (JFE
  2005) Cole, Goldberg White (JFQA 2004)
• Banks spatially price loans
• Petersen Rajan (JF 2002) Degryse Ongena (JF
  2005) Bharath, Dahiya, Saunders Srinivasan
  (JFE 2006), Agarwal Hauswald (2006)

Banks internal organization matters for branch
reach and spatial pricing of loans Also rival
banksorganization matters This paper!
3
Objective Information Travels Better than
Subjective Information
Transportation Costs Differ Number of visits
or mode of communication By borrower to bank By
bank to borrower may differ if banks
organization implies different types of
information Subjective or Soft information more
visits Objective or Hard information fewer or no
visits
4
Transportation Cost Model stylized version with
linear transportation costs

• Banks A, B located at endpoints of line (length
  1)
• Transportation costs ti per unit of distance,
  shaped by internal organization of bank i
• Consider borrower at location x
• Cost visiting A tAx
• Cost visiting B tB(1-x)
• Cost taking loan at A rAx tAx
• Cost taking loan at B rBx tB(1-x)

5
Equal Linear Transportation Costs
Now introduction different costs, but for
graphical purposes, we assume tA t.
Loan Rate

t
-2t
MC 0
0
1/2
1
Bank A
Bank B
Distance
6
Drop in tB (e.g., B employs more objective
information)As reach shrinks and spatial
pricing becomes softer
Loan Rate

tA
tB
tA tB
- (tA tB)
MC 0
0
1/2
tB / (tA tB)
1
Bank A
Bank B
Distance
7
In the paper, we also model

• Non-Linear transportation costs
• Banks with different marginal costs
• Soft Firm Hard Firm and Soft Bank Hard
  Bank
• Soft firms generate more soft information
• If bank A has a comparative advantage in soft
  information
• Bank As reach decreases when B becomes harder
• But less so for soft firms

8
Data We combine two data sets
1

• Portfolio from one Belgian bank
• 17,776 loans to 13,104 borrowers in August 1997
• Degryse Van Cayseele (JFI 2000) Degryse
  Ongena (JF 2005)
• sole proprietorships (81), small, medium, and
  large firms
• Loan Characteristics
• Origination Date, Loan Rate, Collateral,
  Repayment Duration, Purpose, Other
• Relationship Characteristics
• Main Bank and Duration
• Firm Characteristics and Identity (incl. Address)

9
Postal Zone

837 Postal Zones 7,477 Bank Branches
Lender
also possible
Competitors
6 km
Borrower
6 km
10
Distance

• shortest traveling time, in minutes
• to Lender
• to Closest quartile (Bank) Competitor in the
  borrowers postal zone
• 17,776 293,170 borrower - bank branch
  combinations
• Recording errors 801 at closing branches 1 -
  screen (postal zone check)
• 612 contracts in postal zones without competitors
• 15,044 remaining contracts

11
Dependent Variables
12
2

• Bank organization variables
• (source Belgian Bankers Association)

13
Other Control Variables
14
Other Control Variables (contd)
15
Table 2 Impact on Branch Reach
16
Table 4 Spatial Pricing
17
Differential Effects work in progress

• Effect of hierarchy on branch reach is smaller
  for small firms with limited liability
• Effect of hierarchy on spatial pricing is larger
  for small firms with limited liability
• Not incorporated in paper yet
• Layers to Telex softens spatial pricing
• Span of control little impact

18
Conclusions

• Simple model shows that
• Branch reach and severity of spatial loan pricing
  depend on the organization of competing banks
• Empirical Tests
• Branch reach shrinks
• Rival bank is large, hierarchical, or has fax
• Spatial pricing softens
• Rival bank is large or hierarchical