does uncertainty matter for loan charge-offs? laetitia lepetit, frank strobel and david dickinson...
TRANSCRIPT
Does uncertainty matter for loan charge-offs?
Laetitia Lepetit, Frank Strobel and David Dickinson
Research Conference on Safety and Efficiency of the Financial System, Manila
August 27, 2007
2
Objective
• analysis of loan charge-off behavior of banks– non-performing loans need to be charged off
– behavior has discretionary elements, e.g. management of regulatory capital
– consider additional element related to uncertainty about economic environment
• build theoretical model and perform empirical analysis
3
Theoretical model
• develop stylized real options model of bank's decision whether or not to charge off a currently non-performing loan when– there is chance of full loan recovery
– expected present value of reinvesting potentially reclaimed collateral is uncertain
• demonstrate that bank proves less willing to charge off non-performing loan than in classical certainty case
4
Theoretical model (cont.)
• increased uncertainty about and trend in expected value of collateral increases bank's tendency to delay loan charge-off
• better general economic conditions, increasing the probability of loan recovery, reduce bank's willingness to delay loan charge-off
banks have discretionary behavior regarding loan charge-offs related to variability in market for collateral and general economic conditions
5
Empirical analysis
• test this hypothesis of "uncertainty management" in loan charge-offs
• use a panel of European commercial, savings and cooperative banks over the period 1992 to 2005
• estimate the following equation using GMM
1 1 2 3
4
log( ) 3it it it it it it
it i it
NCO NCO X GDP R M M
H
6
Eq.
(Obs) it 1
( )
NCO
it
( )
NPL
it 1
( )
LOANS
it
( )
TCR
it
(?)
LOG(TA) it
( )
M
it
( )
H
it
( )
log(GDP)
it
( )
R3M
J-stat (P-
value)
Arellano and Bover (1995) estimator
(2.1) (552)
0.090*** (140.651)
0.063*** (42.169)
0.004*** (42.413)
_ _ -0.014*** (-103.057)
-0.006*** (-7.792)
0.067*** (36.479)
0.032*** (38.980)
41.489 (0.45)
(2.2) (552)
0.111*** (137.582)
0.062*** (68.586)
0.004*** (33.238)
_ 0.562*** (72.010)
-0.018*** (-79.389)
-0.009*** (-11.043)
0.095*** (29.067)
0.070*** (72.798)
41.781 (0.35)
(2.3) (506)
0.091*** (4.467)
0.065*** (5.151)
_ 0.014 (1.048)
_ -0.006*** (-3.955)
-0.002* (-1.814)
0.036*** (2.593)
0.062*** (7.397)
30.405 (0.44)
Empirical analysis (cont.)