casmef seminars discussion of : the recapitalization of banking and insurance during the 2007-09...

5
CASMEF seminars Discussion of : The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis Zeno Rotondi – Head of UniCredit Research Italy May 13, 2011 – LUISS University, Rome

Upload: alvin-copeland

Post on 01-Jan-2016

217 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: CASMEF seminars Discussion of : The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis Zeno Rotondi – Head of UniCredit Research

CASMEF seminars

Discussion of :

The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis

Zeno Rotondi – Head of UniCredit Research Italy

May 13, 2011 – LUISS University, Rome

Page 2: CASMEF seminars Discussion of : The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis Zeno Rotondi – Head of UniCredit Research

2

Main issue / background literature / findings

1. What are the determinants of banks and insurers recapitalizations during the subprime crisis of 2007-09?

2. No micro-data based empirical analyses (at the moment)

3. The intensity of recapitalizations is related to: exposures to toxic assets

probability of distress (Merton probability of default, MPD)

funding risk

Page 3: CASMEF seminars Discussion of : The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis Zeno Rotondi – Head of UniCredit Research

3

Methodology

total obs 97

binomial logit (no capital increase; private market recap/bailout)

ordered logit (no capital increase; private market recap; bailout)

further ordered logits…..up to 5 states ordering

three phases of the crisis are examined: september 2007 (Northern Rock); april 2008 (Bear Stearns); september 2008 (Lehman)

Explanatory variables in the baseline equation:

MPD = Merton probability of default

LIQ = dummy variable which captures liabilities maturity

TOXIC_1 = high toxic asset exposure / tangible common equity

TOXIC_2 = all toxic asset exposure / tangible common equity

No control variables (?)

Page 4: CASMEF seminars Discussion of : The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis Zeno Rotondi – Head of UniCredit Research

4

Comments (1/2)

sample small and mixed (different explanatory variables for banks and insurers recap?)

add control variables: type of institution; bank specialization; insurer specialization; country fixed effects; size (log of total assets)

add bank specific regressors (pre-crisis values of bank characteristics, for ex. 2006 values): deposits scaled by total assets; loan loss provisions scaled by total assets;non-performing loans scaled by total gross loans; etc.

what about the intensity of recap?: the recap dummy is equal to one if a bank received at least one capital injection between september 2007- march 2009 and zero otherwise. Recap size implies considering the sum of all capital injections received during this period

Page 5: CASMEF seminars Discussion of : The Recapitalization of Banking and Insurance during the 2007-09 Credit Crisis Zeno Rotondi – Head of UniCredit Research

5

Comments (2/2)

relevant omitted explanatory variable: Tier1 risk-weighted capital ratio

it is not examined whether more Tier1 capital reduces the probability of being recapitalized: in joint consideration with the fact that many banks appeared to be in compliance with regulatory capital requirements before the crisis (see for example Demirgüç-Kunt et al. 2010) this outcome would be relevant for regulators (i.e. augment the central role of capital requirements compared to the crisis period)

Using the risk-weighted Tier1 capital ratio may lead to identify total balance sheet size as an important predictor of recapitalizations, alternative/additional to the indicators of early warning of financial distress considered in the paper