financial crises

Report
Econ 492:
Comparative Financial Crises
Lecture 1
12 September 2012
David Longworth
This material is copyrighted and is for the sole use of students registered in ECON 492. This material
shall not be distributed or disseminated to anyone other than students registered in ECON 492. Failure
to abide by these conditions is a breach of copyright, and may also constitute a breach of academic
integrity under the University Senate’s Academic Integrity Policy Statement.
Overview
I.
II.
III.
IV.
Introduction
Financial System: Overview
Causes of Financial Crises
Prediction of Financial Crises
Note: AG indicates Franklin Allen and Douglas Gale
(2009), Understanding Financial Crises. KA indicates
Charles P. Kindleberger and Robert Aliber (2011 or
2005), Manias, Panics, and Crashes.
Economics 492 Lecture 1
2
l. Introduction
• Introductions
Economics 492 Lecture 1
3
l. Introduction
• Introductions
• August 2007
Economics 492 Lecture 1
4
l. Introduction
• Introductions
• August 2007
• Outline of Course (handout)
Economics 492 Lecture 1
5
l. Introduction
• Outline of Course (after Fin. Sector Overview)
(Prediction)
Causes
Transmission
Prevention
Policy
Response
Economics 492 Lecture 1
6
l. Introduction
•
•
•
•
The paper
Presentation and discussion
Participation
Schedule
–
–
–
–
–
–
Lecture next week
26 September: your (two paragraph) topic due
4 October (Thursday): your 2-page outline due
4 – 24/25 October (and beyond): weekly consultations
31 October - 21 November: presentations
28 November: paper due in class and class discussion
Economics 492 Lecture 1
7
II. Financial System: Overview
•
•
•
•
Roles played by the financial system
Bank balance sheets (assets and liabilities)
Risks faced by banks
Market failures
Economics 492 Lecture 1
8
II. Financial System: Overview
• Roles played by the financial system
– Channelling savings into investment/Efficient
allocation over time (consumption/saving,
production) (role played by markets, banks, pension
funds)
– Transferring risk (role played by markets and by banks
and insurance companies)
– Making markets and providing liquidity (and its
various guises and definitions) (role played by markets
and by banks)
– Maturity transformation (role played by banks)
– Effecting payments (role played by banks)
Economics 492 Lecture 1
9
II. Financial System: Overview
• Roles played by the financial system
– Channelling savings into investment/Efficient
allocation over time
Economics 492 Lecture 1
10
II. Financial System: Overview
• Roles played by the financial system
– Transferring Risk
• Risk sharing
– There is always uncertainty about the future, e.g. income
– Contingent commodity: “a good whose delivery is contingent
on the occurrence of a particular state of nature” (AG)
– Two equivalent ways to achieving an efficient allocation of risk
» If there are complete markets for contingent commodities
(consumers only have to satisfy their budget constraint)
» If there are Arrow securities for each state, securities
which are a “promise to deliver one unit of money if a
given state occurs and nothing otherwise.” (AG)
Economics 492 Lecture 1
11
II. Financial System: Overview
• Roles played by the financial system
– Transferring Risk
• Insurance and pooling risk
– When there is a large number of consumers that can be
assumed to be independent, the law of large numbers can be
used to predict the average outcome
– This is what insurance companies do, pooling large numbers
of (largely) independent risks, so that the aggregate outcome
is approximately constant (each individual could be given a
constant level of consumption)
Economics 492 Lecture 1
12
II. Financial System: Overview
• Roles played by the financial system
– Making Markets and Providing Liquidity
• Liquidity definitions
– Assets are liquid “if they can easily be converted into
consumption without loss of value.” (AG)
– Consumers have a preference for liquidity to the extent that
they have uncertainty about the timing of their consumption
» therefore want to hold liquid assets.
• Financial institutions (banks) can rely on law of large
numbers to provide liquidity to consumers while
investing some of their savings in illiquid assets
Economics 492 Lecture 1
13
II. Financial System: Overview
• Roles played by the financial system
– Making Markets and Providing Liquidity
• Theory of banking relies on:
– “A theory of liquidity preference, modeled as uncertainty about the
timing of consumption.” (AG)
– “The representation of a bank as an intermediary that provides insurance
to depositors against liquidity (preference) shocks.” (AG)
» Promises of consumption contingent on the date of withdrawal
– “A maturity structure of bank assets, in which less liquid assets earn
higher returns.” (AG)
» Banks have relatively liquid liabilities and relatively illiquid assets.
“They borrow short and lend long.” (AG)
• Asset markets and market makers provide liquidity to agents who
may be holding otherwise illiquid assets (stock and bond markets)
Economics 492 Lecture 1
14
II. Financial System: Overview
• Roles played by the financial system
– Maturity transformation (role played by banks)
• As we have already seen, banks tend to have shorterterm liabilities than the maturity of their assets
– Effecting Payments (role played by banks)
• This is another reason why consumers place deposits
with banks. It is extremely difficult for consumers not to
transact with banks because of this. Currency and
demand deposits are close to perfect substitutes, but it
is difficult to make large payments with currency.
Economics 492 Lecture 1
15
II. Financial System: Overview
• Bank balance sheets (assets and liabilities)
Assets
Liabilities
Non-mortgage loans
Short-term retail deposits
Mortgage loans
Long-term retail deposits
Risk-free bonds
Short-term wholesale deposits
Short-term risky bonds
Bonds
Long-term risky bonds
Foreign currency liabilities
Derivatives and repos
Derivatives and repos
Foreign currency assets
Equity (capital)
Total Assets
equals Total Liabilities
Economics 492 Lecture 1
16
II. Financial System: Overview
• Canadian banks: balance sheet(May ‘12, $bn.)
Assets
Liabilities
728.3 Non-mortgage loans
743.7 Demand & notice deposits
881.9 Mortgage loans
573.6 Other (typically long-term) deposits
252.9 Risk-free bonds (government)
594.3 Subordinate Debt & other liabilities
181.1 Risky bonds & stocks
174.3 Other (including derivatives)
1573.2 Foreign currency assets
62.5 Other
1620.2 Foreign currency liabilities
197.4 Shareholder equity
3791.7 Total Assets
equals
3791.7 Total Liabilities
Economics 492 Lecture 1
17
II. Financial System: Overview
• Risks faced by banks:
– credit risk (key to beginning of crises)
– funding liquidity risk (key to transmission of crises)
– market liquidity risk
– market risk
– risk to the value of collateral (including housing)
– exchange rate risk (if not matched; important in
EMEs)
Economics 492 Lecture 1
18
II. Financial System: Overview
• Risks faced by banks:
– credit risk (CR): risk of loss due to a debtor’s nonpayment of a loan or a bond or a derivative
• Applies to everything on asset side of balance sheet
except risk-free bonds
Economics 492 Lecture 1
19
II. Financial System: Overview
• Risks faced by banks:
– funding liquidity risk (FLR): risk that over a specific
horizon a bank will be unable to settle its
obligations with immediacy (from Drehmann and
Nikolaou, 2009)
• Applies to risk of not having the ability to raise deposits
or bond or equity funding (liability side of balance
sheet) or to sell off highly liquid assets (such as risk-free
assets) when required to meet obligations
• Note the obvious connection to bank runs
Economics 492 Lecture 1
20
II. Financial System: Overview
• Risks faced by banks:
– market liquidity risk (MLR): risk that the market
will not trade an asset (as in a financial crisis), or
not trade it near the pre-existing price
• Banks face this risk on risky bonds and on derivatives
• There are various measures of asset liquidity
– Bid-offer spread: a measure of transactions cost
– Market depth: amount that can be transacted at various
spreads
– Immediacy: time needed to trade a certain amount at a price
– Resilience: speed at which prices return to pre-existing levels
Economics 492 Lecture 1
21
II. Financial System: Overview
• Risks faced by banks:
– market risk (MR): risk that the value of a portfolio
will change due to changes in overall market risk
factors (stock prices, interest rates, foreign
exchange rates, and commodity prices). This is
systematic risk that cannot be diversified.
• Banks face this on their portfolios of bonds and
derivatives
• Particularly important for categories that need to be
marked-to-market for accounting purposes (trading
book) because they are not going to be held to maturity
Economics 492 Lecture 1
22
II. Financial System: Overview
• Risks faced by banks:
– value of collateral risk (VCR) (including housing and
securities): risk that the collateral that backs loans
that banks have made will change, giving the bank
more exposure to credit risk.
• Banks face this on mortgage loans (if uninsured), car loans,
derivative products and repos
– exchange rate risk (ERR): risk that a bank’s value will
change when the amount of its foreign currency
assets differ from the amount of its foreign currency
liabilities (in a given foreign currency) adjusted for
derivative positions relevant for covering this risk
Economics 492 Lecture 1
23
II. Financial System: Overview
Assets
Liabilities
Non-mortgage loans (CR)
Short-term retail deposits (FLR)
Mortgage loans (VCR, CR)
Long-term retail deposits (FLR at
maturity)
Risk-free bonds (MR) (-FLR)
Short-term wholesale deposits (FLR)
Short-term risky bonds (MR, MLR,
CR)
Bonds (FLR at maturity)
Long-term risky bonds (MR, MLR,
CR)
Foreign currency liabilities (ERR,
other)
Derivatives and repos (MR, MLR,
VCR, CR)
Derivatives and repos (MR)
Foreign currency assets (ERR , other)
Equity (capital)
Economics 492 Lecture 1
24
II. Financial System: Overview
• Market failures (& justification for regulation):
– Failure of bank will lead to a loss of access to future credit for
bank’s small and medium-sized customers
– Failure or severe weakening of several banks will likely lead to a
credit crunch and loss of access
– Informational contagion can arise if banks are seen to have
similar assets or funding models
– Interconnection of banks means that failure of one bank will
lead to uncertainty about other banks
– When liquidity problems are widespread, liquidity-marginleverage cycles can arise, as well as fire sales
– In boom phase excessive credit expansion can lead to resource
misallocation
– (Source: Brunnermeier et al. (2009))
Economics 492 Lecture 1
25
III. Causes of Crises
• Types of financial crises:
– Banking crisis (financial sector crisis)
– Currency crisis
• Forced depreciation
• Large devaluation
– Sovereign debt crisis
• Domestic debt
• External debt
• Most of lecture time will deal with banking
crises
Economics 492 Lecture 1
26
III. Causes of Crises
Asset price
bubbles
Innovation, deregulation,
poor risk assessment,
asymmetric payoffs, poor
government policy
Rapid credit
growth
Financial crisis
and
transmission
Easy monetary or
exchange rate
policy
Economics 492 Lecture 1
27
III. Causes of Crises
• Rapid growth in credit (aggregate or sector
specific) seems to be present in almost all
financial crises, especially the ones associated
with banking crises or asset price bubbles
– Kindleberger and Aliber: “For historians each
event is unique. In contrast economists maintain
that there are patterns in the data and particular
events are likely to introduce similar responses.
History is particular; economics is general.”
Economics 492 Lecture 1
28
III. Causes of Crises
• Minsky focused on procyclical changes in the
supply of credit
– In the expansion, investors gain optimism, raise
estimates of profitability of investments, borrow
more
• Lenders also gain optimism, lowering risk assessments
and becoming less risk averse, and lend more
– In the contraction, investors lose their optimism,
lower estimates of profitability, borrow less
• Lenders have increased loan losses and become more
cautious, lending less
Economics 492 Lecture 1
29
III. Causes of Crises
• Minsky believed that the procyclical changes
in the supply of credit led to financial fragility
and an increased probability of a crisis (KA,
Ch. 3)
– Cycle started by a displacement: an exogenous
outside shock
– This displacement leads people to believe there
are improved profit opportunities in at least one
important sector of the economy. Borrowing rises.
– This expansion of credit fuels the boom
Economics 492 Lecture 1
30
III. Causes of Crises
• Minsky
– Euphoria might develop: investors buy in
expectation of capital gains
• Loan losses incurred by the lenders decline and they
respond and become more optimistic and reduce the
minimum downpayments and the minimum margin
requirements
• There can be a move away from normal rational
behaviour to manias or bubbles
Economics 492 Lecture 1
31
III. Causes of Crises
Displacement
Expected Profit
Opportunities
Mania or bubble
(eventually
ends)
Economics 492 Lecture 1
Borrowing
increases:
fuels boom
Expected capital
gains lead
investors to buy
32
III. Causes of Crises
Hedge
finance
Speculative
finance
Which stage one is at
depends whether
both principle and
interest are being
paid, just interest, or
not even interest.
Ponzi
finance
Economics 492 Lecture 1
33
III. Causes of Crises
• The rapid growth in credit (aggregate or sector specific) can
arise from a number of factors:
– Financial deregulation/liberalization (KA, Reinhart)
• Ex.: Led to bubbles in real estate and stocks in Nordic countries
–
–
–
–
Japanese banking regulators had eased restrictions on Japanese banks abroad
Nordic banking regulators had eased restrictions on banks borrowing abroad
Loans to Nordic borrowers fueled the bubbles
Banks failed
– Financial innovation
• Examples in the most recent crisis
–
–
–
–
–
Sub-prime loans
Adjustable rate mortgages in the United States (with teaser rates)
Asset-backed Commercial Paper (with unspecified assets backing it) in Canada
CDOs: collateralized debt obligations
SIVs
Economics 492 Lecture 1
34
III. Causes of Crises
• The rapid growth in credit can arise from:
– Poor assessment of risk (risk management, black swans)
• Value-at-risk calculations were based on short historical periods that did
not include “stressed periods” of abnormally high volatility
• These and other calculations were often based on the “normal” statistical
distribution, but the empirical distributions were often fat-tailed
• Black-swan phenomena: because they haven’t been seen (like black
swans, which were native only to Australia), they will never be seen
• Correlations tend towards one in crises, so the benefits from
diversification tend to disappear then
• Generally, an overemphasis on statistical models (sometimes, overly
simple ones), with not enough emphasis on thinking about what could go
wrong with a given portfolio, especially where there had been financial
innovations
– Movements in interest-rate spreads (over governments) on new products don’t say much
• Overemphasis on rationality of markets
Economics 492 Lecture 1
35
III. Causes of Crises
• The rapid growth in credit can arise from:
– Too big to fail
• Managers and shareholders believe that their bank is
too-big-to-fail and that they will be bailed out by the
government if they are in danger of failing. Therefore
they take more risks.
• Bondholders and large depositors believe that they will
be bailed out by the government and are therefore
willing to lend more to banks at narrow spreads over
risk-free rates even when the banks are expanding
credit rapidly
Economics 492 Lecture 1
36
III. Causes of Crises
• The rapid growth in credit can arise from:
– Compensation schemes for management and
traders that are asymmetric and not risk-adjusted
– Compensation is based on business volume and current
profits, even when investments are longer-term.
– CEOs and traders may have left (or may leave) when the
chickens come home to roost.
Economics 492 Lecture 1
37
III. Causes of Crises
• The rapid growth in credit can arise from:
– Inappropriate government policy regarding
sectoral credit expansion
• e.g., encouraging loans to those who are unable to pay
them
– Example, U.S. sub-prime
Economics 492 Lecture 1
38
III. Causes of Crises
• The rapid growth in credit (aggregate or sector
specific) can also arise from easy monetary
policy or inappropriate exchange rate policy
(especially fixed exchange rates)
– Argument that U.S. monetary policy was too easy
from 2003-07
• Too focused on dangers of deflation and too little
focused on special factors keeping down inflation
Economics 492 Lecture 1
39
III. Causes of Crises
• Note: rapid growth in credit raises leverage,
which is the ratio of assets to capital (net
worth)
• Leverage is closely related to the following
debt ratios:
– (for households): debt / personal dispos. income
– (for businesses): debt / equity
– (for governments): debt / GDP
Economics 492 Lecture 1
40
III. Causes of Crises
• The Austrian school (e.g., Hayek) puts
emphasis on the excess capital that is created
during the credit boom. This inhibits
investment during the recovery period.
Economics 492 Lecture 1
41
III. Causes of Crises
• Credit expansion places the emphasis on credit
decisions gone bad and the eventual insolvency
(failure)of banks.
• When we talk about transmission, we will see
that we can’t entirely divorce insolvency from
illiquidity.
– Banks can become illiquid even if they are solvent
– In turn, their illiquidity can lead to their insolvency
– Suspicion about solvency, however, is the greatest
cause of illiquidity
Economics 492 Lecture 1
42
III. Causes of Crises
• Asset bubbles are typically a contributing factor to
crises. They typically are supported by rapid sectoral
credit expansion.
– When asset bubbles are not associated with rapid credit
expansion, there is typically less of a problem when they
burst because the financial sector is less harmed (banks
don’t typically fail in such circumstances).
• Example: Tech bubble in late 1990s
– Asset bubbles most common in housing or real estate, and
in stock markets (but also exchange markets, commodities)
– Shiller has been a proponent of a psychological effect in
the development of bubbles (rejects efficient markets)
Economics 492 Lecture 1
43
III. Causes of Crises
• 7 biggest financial bubbles in last 40 years (KA)
– Bank loans to Mexico and other developing countries in
the 1970s
– Japanese bubble in real estate and stocks in late 1980s
– Nordic bubble in real estate and stocks in late 1980s
(Finland, Norway, Sweden)
– Asian crisis following bubble in real estate and stocks in
Thailand, Malaysia, Indonesia and others
– Capital flows into Mexico in 1990-93
– Tech stock bubble in U.S. from 1995-2000
– Bubbles in real estate in U.S., U.K., Ireland, Iceland, Spain
and Hungary in early 2000s and the world financial crisis
2007-2009(-2012)
Economics 492 Lecture 1
44
III. Causes of Crises
• “Quantitative antecedents of financial crises”
from Carmen Reinhart (NBER, March 2012):
– Capital inflows
– Equity prices
– Housing Prices
– Inverted V real output growth
– Rise in indebtedness (private or public sector)
Economics 492 Lecture 1
45
III. Causes of Crises
• For those who will not be using statistical
methods, “Event Study Graphs” (e.g., IMF
GFSR, September, 2011, Chapter 3, p.11) can
be helpful
– Plot key indicators (e.g., credit/GDP, real asset
prices) from T-k through T (onset of crisis) to T+k
– Could average a number of crisis countries (actual
dates of T could differ across countries)
– Could compare to an average of paired non-crisis
countries
Economics 492 Lecture 1
46
III. Causes of Crises
Credit/GDP
8
7
6
5
4
Credit/GDP
3
2
1
0
T-3
T-2
T-1
T
T+1
Economics 492 Lecture 1
T+2
T+3
47
III. Causes of Crises
There are often
connections among the
three common types of
“financial” crisis.
Financial Sector
Crisis
Exchange Rate
Sovereign Debt
Crisis
Crisis
Economics 492 Lecture 1
48
III. Causes of Crises
Simultaneous Crises
(within one year):
Laeven & Valencia (2012)
Banking Crises
99
28
Currency
Crises
153
11
8
29
Debt Crises
18
Economics 492 Lecture 1
49
III. Causes of Crises
Beginning of
Banking Crisis
Currency Crash
Peak of banking
crisis (if no debt
default)
Peak of banking
crisis (if debt
default occurs)
Default on
external or
internal debt
C. Reinhart’s (2012) “The sequencing of crises: A prototype”
Economics 492 Lecture 1
50
III. Causes of Crises
• Sequencing of Crises (Laeven & Valencia,2012)
– 16% of banking crises preceded by currency crisis
(within 3 years)
– 21% of banking crises precede a currency crisis
– 1% of banking crises preceded by sovereign debt
crisis
– 5% of banking crises precede a sovereign debt
crisis
Economics 492 Lecture 1
51
III. Causes of Crises
• Some suggested areas for papers
• What was the relationship between credit growth
(aggregate or sectoral) and various episodes of asset
price bubbles? (In recent crisis, could compare across
US, UK, Spain, Ireland, Iceland, Hungary)
• What has been the relationship between capital flows
and various episodes of asset price bubbles (especially
in EME stock markets or housing markets)?
• What has been the relationship between the sovereign
debt crises in Greece, Ireland, and Portugal (and feared
crises in Spain and Italy) and other types of “financial
crises” in the same countries?
Economics 492 Lecture 1
52
III. Causes of Crises
• Some suggested areas for papers
• Were financial crises worse in countries where the capital
stock grew more rapidly? Put another way, which of the
following predict the severity of a crisis: credit, asset price
bubbles, capital stock growth? Focusing on those crises
where there was a real estate bubble, which of the following
predict the severity of a crisis: mortgage credit, total credit,
real house prices, housing stock growth?
• What are the similarities of the countries in crisis in the Euro
area and how do these compare to the countries not in crisis
in that area? (Event study graphs—what are the interesting
indicators?)
Economics 492 Lecture 1
53
IV. Prediction of Financial Crises
• In the literature, there are two main types of
approaches to prediction
– One deals with Type I and Type 2 errors, and
minimization of noise-to-signal ratio
• A related approach has decision trees
– Second deals with estimation of multivariate logit
(or probit) models (with maximum likelihood
methods)
Economics 492 Lecture 1
54
IV. Prediction of Financial Crises
• Type I and Type II errors and noise-to-signal
ratio
– Type I error is a false positive (no crisis predicted
when was a crisis)
• The significance level α of a test
– Type II error is a false negative (crisis occurred
when none was predicted)
• β, or 1 minus the power of the test
Economics 492 Lecture 1
55
IV. Prediction of Financial Crises
• Type I and Type II errors and noise-to-signal
ratio
– Noise-to-signal ratio: ratio of
• Number of predictions of crisis when no crisis /
number of observations of no crisis, to
• Number of predictions of crisis when crisis /
number of observations of a crisis
– Typically choose rule minimizing noise-to-signal ratio in this
literature
Economics 492 Lecture 1
56
IV. Prediction of Financial Crises
• Type I, Type II errors and noise-to-signal ratio
– Some papers report unconditional probability of a
crisis (c/nT) where c is number of crises, n is
number of countries, and T is number of years
• These papers then go on to report conditional
probabilities of a crisis given the condition A ≥ A* for
some indicator variable A (say, credit growth or level)
• Also, can have a more complicated formulation such as
the conditional probability:
– P(Crisis| A ≥ A*, B ≥ B*, C ≥ C*)
Economics 492 Lecture 1
57
IV. Prediction of Financial Crises
• Type I, Type II errors and noise-to-signal ratio
– Note sufficient condition P (Crisis| A ≥ A’) = 1
– Note necessary condition P (A ≥ A ’’|Crisis) = 1
– Then want to constrain grid search to A’’ ≤ A* ≤ A’
– Note can do in-sample, out-of-sample exercise
• Example, estimate before recent crisis, see how well it
predicts which countries experienced recent crisis
Economics 492 Lecture 1
58
IV. Prediction of Financial Crises
• Type I and Type II errors and noise-to-signal ratio
– Three key articles on prediction of banking crises are Borio
and Lowe (2002 BIS WP) and Borio and Drehmann (2009
BIS WP 284, 2009 March BIS Quarterly Review)
– Important inputs into their work were the dating of
financial crises by Bordo et al. (2001 Appendix A) and work
by Kaminsky and Reinhart on twin crises (1999
AER)(important in its own right)
– Authors use gaps between levels of variables and trends
for those variables, using recursive Hodrick-Prescott
procedure
• Procedure minimizes weighted average of sum of squared
deviations of actual from trend and sum of squared second
differences of trend
Economics 492 Lecture 1
59
IV. Prediction of Financial Crises
• Type I, Type II errors and noise-to-signal ratio
– The main result from the Borio-Drehmann piece is
that, when credit-to-GDP ratios relative to trend
and real property prices relative to trend and/or
real stock prices relative to trend are quite high,
the probability of a banking crisis is quite
elevated.
Economics 492 Lecture 1
60
IV. Prediction of Financial Crises
Horizon
(Years)
In Sample
% Predicted
In Sample
Noise/Signal
Out of Sample
% Predicted
Out of Sample
Noise/Signal
1
46
0.23
67
0.53
2
69
0.13
67
0.53
3
69
0.11
67
0.53
Borio and Drehmann Model is (property gap greater than 15 or equity gap greater
than 60) and credit gap greater than 6.
Economics 492 Lecture 1
61
IV. Prediction of Financial Crises
• Logit (or probit) model
– Logit model is based on logistic function:
•
•
•
•
•
  = 1/(1 +  − )
Where  = β0 + β1 1 + … + β 
Note, as  → −∞ ,   → 0 (no crisis)
Note , as  → ∞, () → 1 (crisis)
If larger values of x are associated with a greater
chance of crisis then their coefficients will be positive.
Economics 492 Lecture 1
62
IV. Prediction of Financial Crises
• Logit (or probit) model
– Probit model is based on the inverse cumulative
density function of the standard normal
distribution
– (Many feel that practically the differences
between probit and logit are not too large.
Estimates of multivariate logit tend to converge
better/faster than estimates of multivariate
probit.)
Economics 492 Lecture 1
63
IV. Prediction of Financial Crises
• Logit Regressions
– Schularick and Taylor (2009 NBER) show that it is real loan
growth, not real money growth, than predicts banking
crises historically
– Reinhart-Rogoff (2011 AER) show that This time is Not
different in many different dimensions.
• Systemic banking crises in financial centers (U.S., U.K., and
historically France) help explain domestic banking crises
• External-debt-to-GDP ratio helps explain banking crises
• Domestic banking crises help explain sovereign debt defaults
• Growth in public debt helps explain sovereign debt defaults
Economics 492 Lecture 1
64
IV. Prediction of Financial Crises
• Logit Regressions
– Gourinchas & Obstfeld (2012, AEJMacro) show
• Domestic credit growth and real currency appreciation
are best predictors of crises in general (advanced
countries and EMEs)
• Higher level of foreign exchange reserves reduces
probability of crises
• Public indebtedness can raise probability of banking
(and currency) crises in EMEs
– Eichengreen (2004, Ch.5) estimates regressions for
various exchange rate “events”, including “crises”
Economics 492 Lecture 1
65
IV. Prediction of Financial Crises
• Probit Regressions
– IMF (GFSR, Sep. 2011) shows that higher credit-toGDP gap or credit-to-GDP growth will increase
probability of systemic banking crisis
• Probabilities increase further when lagged increase in
equity prices is higher
Economics 492 Lecture 1
66
IV. Prediction of Financial Crises
• Some suggested areas for papers:
– Add to Borio and Drehmann an international
dimension: contagion from other crises or assets of
domestic banks in foreign countries or …
– In Reinhart and Rogoff (2011) explore whether the
inclusion of both external debt ratio and growth in
overall public debt changes conclusions about what
predicts banking crises and defaults post 1974
– Using estimated equations from various authors, what
are the probabilities of sovereign defaults by Spain
and Italy and other countries with high or rising debt
at present?
Economics 492 Lecture 1
67
IV. Prediction of Financial Crises
• Some suggested areas for papers:
– What predicts the spread in long-term rates
between the periphery countries in the euro area
and Germany? (events and longer-term factors)
– To what extent does net government debt add
information to gross government debt in
predicting sovereign debt crises?
– Add real house price gaps and real equity price
gaps to specifications of Gourinchas and Obstfeld
Economics 492 Lecture 1
68
Suggestions for this week
• Review Lecture at ECON 492 web site
• Skim through Reference List for ECON 492 on ECON 492
web site; what sparks your interest? Pursue a reference.
• Read one or two of:
– KA chapters 1-3 (skim the rest)
– Reinhart and Rogoff chapters 1, 10, 13 (skim the rest)
– Gourinchas & Obstfeld (2012, AEJMacro) “Stories of the 20th
Century for the 21st”
– Reinhart and Rogoff (2011, AER), “From Financial Crash to Debt
Crisis”
– The Squam Lake Report chapter 1
• If you have an idea of what you would like to do already,
see me today or tomorrow, or e-mail me
Economics 492 Lecture 1
69

similar documents