Diapositiva 1

1
Discussion of A Model of Shadow Banking (by N.
Gennaioli, A. Shleifer and R. Vishny)
Oscar Arce CNMV - Research and Statistics
Department Director
Debt and Credit, Growth and Crisis BdE -
WB Madrid, 18 June 2012
2

What the paper does

• It provides a comprehensive story on the links
  between aggregate savings (the savings glut),
  the scarcity of safe assets and the role of
  securitization in their production, interest
  rates, leverage and financial stability.
• A relatively simple model
• - A demand for safe assets by risk-averse
  investors (lenders)
• - A supply of safe assets by risk-neutral
  intermediaries. Two cases
• i) If low demand, intermediaries own
  collateral suffice to meet it
• ii) If high demand, intermediaries produce
  new safe assets securitization
• 3. Central results
• If risks are correctly assessed, securitisation
  contributes to financial stability (risk
  diversification -pooling) and is Pareto-improving
  (efficient risk allocation tranching and
  originators retention).
• If tail risks are neglected, securitisation may
  become a driver of financial instability through
  ex post 1) excessive leverage and 2) illiquidity

3

What I like most

• A relatively simple model sheds light on the
  links between some key developments before and
  during the crisis in a credible way
• The shocks-transmission mechanisms (aggregate
  contagion of idiosyncratic risks through
  securitization and illiquidity) help understand
  the velocity and intensity with which the initial
  turbulences of August 2007 spread over the entire
  system
• Some key pieces in the model are well grounded on
  empirical evidence
• The role of securitization in producing
  high-quality assets
• The role of securitization holdings by leveraged
  intermediaries in amplifying the shocks

4

Key model elements (I)

• The role of securitization in producing
  high-quality assets

Global issuance of AAA-rated long-term fixed
income assets (1990-2009)
Source Dealogic
The rapid growth in global AAA-assets (from 20
of total assets in 1990 to 60 in 2003) was
largely due to securitisation
5

Key model elements (II)

• The high volume of securitization in the
  leveraged institutions balance-sheets (key for
  the transmission of shocks in the model and in
  reality too)

Distribution of the holdings of GSE-backed
securities
Source Shin (2009)
Leveraged institutions (commercial banks,
investment banks and hedge funds) were the main
holders of ABS
6

What I find less convincing

• A static and partial equilibrium framework is
  used to explain some facts of general
  equilibrium (systemic) nature, thus missing some
  interesting dimensions, e.g.
• In the model only one type of productive
  investment which exogenous wrt to how it is
  financed.
• In reality, most problematic ABS were linked to
  mortgages (RMBS and CMBS). Some argue that the
  housing bubble was indeed fuelled by
  securitisation
• A conjecture if securitisation favours the
  emergence of housing bubbles, then securitisation
  may not be that good in the first place, even
  under RE (e.g. in terms of financial stability or
  general welfare e.g. Arce López-Salido 2011)
• A (big) conjecture can too much securitisation
  generate endogenous aggregate instability through
  its effects on the investment mix?

7

What I find less convincing

• The assumption that originators (think that)
  retain all the risk seems at odds with data
  ----too little skin in the game

Tranche retention by originators ( average share
of retained deal volume)
Source IMF GFSR

• More realistic to assume two types of
  intermediaries
• i) some with productive-investment opportunities
  that sell ABS
• ii) some without such opportunities that buy ABS
  and not only their AAA- tranches
• - Also, relax the assumption of ex ante symmetric
  information would be interesting

8

What I find less convincing

• In the model, the positive link between
  securitisation and leverage depends on the volume
  of retained tranches.
• But, if retention is low (O2D model), that link
  weakens ? in the limit, originators create and
  sell assets with no impact on the size of their
  balance-sheets.
• On the other hand, investors in ABS may well end
  up raising their leverage.

Bank Leverage
Distribution of the outstanding stock of ABS

• (Net) originators seems less leveraged than (net)
  investors

9

What I find less convincing

• Is the papers liquidity contagion channel the
  most relevant??
• An alternative based on asymmetric information
  would help us understand the collapse of the
  secondary market of European securitisations,
  whose performance has been remarkably good (i.e.
  a massive lemons problem)

Rating migrations Spain, Europe and U.S

• Asset complexity, lack of transparency and a
  minimum degree of standardization have
  exacerbated the asymmetric information, killing
  the EU secondary market

10

Conclusions

• Great paper, it helps understand some key links
  in the financial system that are crucial in the
  origin, the propagation and the depth of the
  crisis.
• Some assumptions and modelling choices are not
  fully convincing but the overall story is a
  credible one.