2. Overview
Part I: Crises are more than ‘volatility’
Part II: Crises are inevitable
Part III: History repeats itself
Part IV: Lessons can be learned
Part V: Takeaways
2
4. I Crises are more than ‘volatility’
Bond yields during the past 5000 years
4
5. I Crises are more than ‘volatility’
Tulip mania (1636-37)
-95%
5
6. I Crises are more than ‘volatility’
South Sea bubble (1720)
Share price of South Sea Company
(in GBP)
-90%
6
7. I Crises are more than ‘volatility’
Germany early 1920s
7
8. I Crises are more than ‘volatility’
US bond yields 197Os
US 10 year bond yields in %
18
16
14
12
10
8
6
4
2
0
8
+13 ppt
9. I Crises are more than ‘volatility’
Belgian bond yields 197Os
Belgian 10 year bond yields
in %
16
14
+9 ppt
12
10
8
6
4
2
0
9
10. I Crises are more than ‘volatility’
US 1929
S&P Composite crash
35
30
25
-85%
20
15
10
5
0
10
11. I Crises are more than ‘volatility’
Effect of the 1929 crash an the Great Depression on Belgium
Belgian unemployment rate
in %
Belgian stock market index
120
25
100
20
80
60
+ 21.6 ppt
-75%
15
10
40
20
0
11
5
0
12. I Crises are more than ‘volatility’
Effect of the 1929 crash an the Great Depression on Belgium
Belgian 10 year
government bond yield
Belgian stock market index
120
(in %)
7
100
80
60
6
-2.6 ppt
-75%
5
40
4
20
0
12
3
13. I Crises are more than ‘volatility’
The Japanese equity and real estate crash (1990)
Japanese equity and real estate
(Jan 1985=100)
600
500
+422 ppt
400
-76 ppt
300
200
100
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
0
Equity (MSCI)
13
Listed real estate (Topix)
14. I Crises are more than ‘volatility’
Effect of Japanses crash in bond yields (1990s)
Japanese 10 year government
bond yield in %
9
8
7
6
5
4
3
2
1
0
14
1998
1996
1994
1992
1990
1988
1986
1984
-7 ppt
15. I Crises are more than ‘volatility’
Financial sector crisis since 2007
EMU Corporate BBB
financial spread to swap in
bps
3000
2500
2000
1500
+24 ppt
1000
500
15
Financials
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
0
16. Overview
Part I: Crises are more than ‘volatility’
Part II: Crises are inevitable
16
18. II.2 Bubbles are possible because of market
irrationality
In efficient markets, irrational exaggerations are highly unlikely
However, grounds for market inefficiency and irrationality include
“The limits of arbitrage” (Shleifer and Vishny, 1997): there are cost of arbitration,
e.g. the fate of LTCM in 1998
Keynes’ “greater fool” game
Psychological biases (Daniel Kahnemann, “Prospect theory”): investors’ utility is
reference based, e.g. on the profits earned by others
Multiple equilibria exist, dependent on expectations. Changes are often
triggered by expectation ‘shifts’ (Kindleberger’s ‘displacements’)
Exaggeration cycles are inherent in market economies
18
19. II.3 Bubbles are possible because of the
‘debt nature’ of our money
Our ‘money’ is a debt certificate of the state or a private economic agent
Coins and bills are debt certificates of the state
They are legal tender,…
… in particular they can be used to settle tax debt towards the state
All other money (deposits, accounts, etc…) are debt instruments of private
agents, mostly banks
Since our monetary system is based on debt and credit, occasional crises
are not avoidable
19
21. II.4 Fractional Reserve Banking
Multi equilibria create a ‘coordination problem’
Pay-off matrix
Person B
Person A and B each have deposits of
100 in the bank
As a result of Fractional Reserve
Banking, the bank has only reserves of
10
Each person has two possible
strategies: to withdraw or not
In the case of no withdrawal, the fact
that the person can use the bank’s
service (safety of deposits and
payments) has an additional monetary
value of 1
There are two stable Nash equilibria
The existence of a lender of last resort
can shift the Nash equilibrium from
bank run to the cooperative
21 equilibrium
Withdraw
P
e
r
s
o
n
Do not
withdraw
Withdraw
(5,5)
(10,0)
Do not
withdraw
(0,10)
(101,101)
A
Two stable Nash equilibria
22. II.4 Fractional Reserve Banking increases
vulnerability to crises
Defining property: not all deposits covered by reserves
This makes the system vulnerable to bank runs
Instability is NOT caused by the nature of fiat money, it applies to gold
standard as well
One way to avoid bank runs would be a system of Full Reserve or 100%
Reserve Banking (Milton Friedman)
Problem: liquidity provision, banking sector cannot play its role of financial
intermediation, leading to credit crunch
Not currently practiced as a system
22
23. II. 5 Saving for the future also requires debt accumulation
Saving in fixed income asset = creating a claim
on future output
But: someone must promise to give up that
part of future output, i.e. incur debt
Implication: saving and debt are two sides of
the same coin: one’s savings are someone else’s
debt
Savings are only possible to the extent that
someone else is prepared to incur debt for the
same amount
If literarily all debts are repaid, all money would
disappear and our monetary system would
23 collapse: back to barter trade
24. Overview
Part I: Crises are more than ‘volatility’
Part II: Crises are inevitable
Part III: History repeats itself
24
25. III.1 Anatomy of crises
Kindleberger-Minsky model
Displacement
Euphoria
Boom
25
Crisis
26. III.1 Anatomy of crises
Kindleberger-Minsky model
Displacement
Euphoria
Boom
26
Crisis
27. III. 1 Anatomy of crises
‘Displacement’ phase
A ‘displacement’ is an long and pervasive exogenous shock to the
macro-economic system that changes expectations and perceived
profit opportunities
Outbreak or end of wars
Widespread adoption
of new inventions
(IT, transportation)
Unexpected change of economic policies, e.g. financial
deregulation or disinflationary monetary policy since the
early 1980s
27
28. III.1 Historically ‘displacements’ can boost
economic growth enormously…
Real world GDP per capita
(1913 = 100)
600
574
500
441
400
300
268
200
138
100
100
29
29
37
44
0
Source: Angus Maddison (2001); IMF; UN
28
29. …especially when supported by globalisation
Lower barriers to exchange and communication
and lower average world import tariffs
Decreasing costs of transport and
communication (in 1990 USD)
250
(for members WTO, in %)
Ocean freight (per ton)
Air transport (per 100 passengermile)
Average import tariff in %
200
Telephone call (3 min. New York
Londen)
150
100
50
0
29
Source: IMF; WTO
30. III.1 Anatomy of crises
Kindleberger-Minsky model
Displacement
Euphoria
Boom
30
Crisis
31. III.1 Anatomy of crises
The ‘boom‘ phase
The changed perception of profit opportunities leads to increased
investment and production
This phase is fuelled by a strong expansion of credit
The expansion of credit is inherently unstable (see also earlier)
Minsky’s ‘Financial Instability Hypothesis’
Credit is unstable and inherently pro-cyclical
31
32. III.1 Anatomy of crises
Kindleberger-Minsky model
Displacement
Euphoria
Boom
32
Crisis
33. III.1 Anatomy of crises
The ‘euphoria‘ phase
Speculative investors appear
Growth is increasingly driven by leverage via the credit channel, ultimately
leading to exaggeration
Three types of investors, with decreasing quality of debt: hedge, speculative
and Ponzi investors
The average quality of debt gradually deteriorates
The boom becomes increasingly debt driven
“There is nothing so disturbing to one’s well being and
judgment as to see a friend get rich.” (Anna Schwartz)
33
34. Diminishing quality of debt
III.1 Types of finance
Hedge
finance
• Only interest can be paid from cash-flow from
investment
Speculative • For capital repayment, the investor relies on new
finance
credit or rolling-over of existing debt
Ponzi finance
34
• Capital and interest can be financed by cash-flow
from investment
• For both capital and interest payments, the investor
relies on capital gains on his aquired asset
35. III.1 Innovation in the financial sector plays an
ambiguous role in the ‘euphoria’ phase
• Warren Buffet (2003) : “Credit default swaps are financial weapons of mass
destruction”
• Paul Volcker (2009): “The only real innovation of the past decades in the
financial industry is the ATM”
35
37. III.1 Anatomy of crises
The ‘crisis‘ phase
The key mechanism leading to the crisis, is the accumulation of debt
of increasingly worse quality
New entrants to speculation are increasingly balanced by insiders who
wish to withdraw
The price of the speculative assets fall and some speculative or Ponzi
investors are unable to repay their loans
Possible triggers include:
the failure of a bank (e.g. Lehman)
the revelation of a swindle (e.g. The original Ponzi scheme)
Sudden realisation that the speculative asset is overpriced (e.g. the
Amsterdam tulips)
Rush to liquidity to ‘liquidate’ the speculative asset and deleverage
Credit crunch: banks cease to lend on the collateral of such assets
37
38. III.1 Anatomy of crises
The ‘crisis‘ phase
Like speculation, the ‘liquidation’ process is feeding on itself
The process stops when
either asset prices have fallen so much, that some investors are willing to
invest in the less liquid asset again;
or trade is cut off or suspended;
or sufficient liquidity is provided to meet the demand for cash
- the need for a ‘lender of last resort’
38
39. III.2 Historical examples
The Amsterdam
Tulip mania
(1636-37)
Displacement
Boom in war
against Spain
Treaty of Utrecht Treaty of
1713: British
Versailles
(slave) trade
with South
America
Speculative
asset
Tulip bulbs,
among other
things
South Sea
Company shares
German FX debt
denominated in
gold
Monetary
expansion
Private credit
Sword Blade
Bank
German central
bank
Lender of last
resort
39
The South Sea
bubble
(1720)
German hyper
inflation
(early 1920s)
None
Bank of England
(since 1694)
none
40. III.2 Historical examples
Wall Street crash Japanese real
(1929)
estate and stock
market crash
Displacement
End of post-war
boom
Economic
expansion phase
Financial
deregulation,
exchange rate
pegs
Speculative
asset
US stocks
Nikkei shares
and land
e.g. real estate,
unsustainable
investments
Monetary
expansion
Stocks bought on from low
margin-calls
interest rate
policy
Bank lending
Lender of last
resort
40
Asian crisis
1997
Federal Reserve
IMF, World Bank,
ADB
Bank of Japan
41. III.2 Historical examples
Sharp rise US
bond yields
(late 1960s and
70s)
EMU Sovereign
debt crisis
(2010-)
Displacement
Transition from
Bretton Woods
system to pure
fiat money
Financial
deregulation,
idea of
‘ownership
society’
Creation of
EMU, leading to
artificially low
interest rates
Speculative
asset
Overly
US real estate
expansive
economic policy
Rising Sovereign
debt
Monetary
expansion
Via current
account deficits
Bank credit,
Originate-anddistribute
model
International
capital markets
Lender of last
resort
41
US sub-prime
crisis
(2008)
Rest of the
world and
Federal Reserve
Federal Reserve
ECB to some
extent (OMTs)
42. Overview
Part I: Crises are more than ‘volatility’
Part II: Crises are inevitable
Part III: History repeats itself
Part IV: Lessons can be learned
42
43. IV.1 Kindleberger Minsky model as early warning
Bitcoin ?
Bitcoin
euphoria 2013
Displacement
Fear currency
debasement by
malicious
governments.
Criminal
opportunities
Speculative
asset
‘Bitcoins’, with
no intrinsic
value nor legal
tender
Monetary
expansion
Private credit
A lot of ‘Ponzi’
investors
Lender of last
43
resort
None
USD per Bitcoin
+1100%
44. IV.1 Kindleberger Minsky model as early warning
Federal Reserve balance sheet ?
Central banks’ balance
sheet expansions since
2008
Displacement
Speculative asset
Fear of new Depression.
This time it’s different:
monetary expansion is noninflationary
Large scale buying of US
Treasuries and Mortgage
Backed Securities = credit
provision. Rollovers are
consistent with Minsky’s
speculative investors
Fed balance sheet
(local currency, Jan 2007=100)
500
450
400
350
300
250
200
150
Monetary
expansion
Credit expansion via
creation of central bank
money
Lender of last
44
resort
Federal Reserve
100
50
Fed
ECB
45. IV.1 Kindleberger Minsky model as early warning
Chinese debt crisis ?
Chinese investment boom
and debt build-up after
2008
Displacement
Speculative asset
Start Quantitative Easing
Federal Reserve in
combination with RMB peg
to USD.
China ‘imports’ US
expansionary monetary
policy
Investment boom financed
by cheap credit
Outstanding amount of private debt
(in % of GDP)
180
US
170
160
150
140
130
120
Monetary
expansion
Credit growth facilitated by
monetary inflow from US
and artificially low interest
rates
Lender of last
45
resort
Chinese central Bank
110
100
EMU
China
46. IV.1 Kindleberger Minsky model as early warning
Chinese debt crisis ?
Chinese investment boom
and debt build-up after
2008
Displacement
Speculative asset
Start Quantitative Easing
Federal Reserve in
combination with RMB peg
to USD.
China ‘imports’ US
expansionary monetary
policy
Investment boom financed
by cheap credit
“Shadow financing” increasingly important
180
Private sector debt in % of GDP
50
Share of new credit other than
bank loans (in %, right)
40
170
160
150
30
140
130
20
120
10
Monetary
expansion
Credit growth facilitated by
monetary inflow from US
and artificially low interest
rates
Lender of last
46
resort
Chinese central Bank
110
100
0
47. IV.1 Kindleberger Minsky as early warning
Emerging Markets: could 1997 crisis happen again ?
Rising external deficits
Emerging Markets since
mid-2000s
External deficits Emerging Markets building up again
(current account balances, in % of GDP)
3
Displacement
Speculative asset
Low global bond yields
(savings glut).
End of commodity and
energy super-cycle
Investment boom
2
1
0
-1
-2
Monetary
expansion
External deficits financed
by inflow of first FDIs than
of portfolio investments
-3
-4
Latin America
Lender of last
resort
47
None (IMF to some extent)
-5
Asia ex China
48. IV.1 Kindleberger Minsky as early warning
Emerging Markets: could 1997 crisis happen again ?
Rising external deficits
Emerging Markets since
mid-2000s
Displacement
Speculative asset
Low global bond yields
(savings glut).
End of commodity and
energy super-cycle
Investment boom
Rising private and public sector debt
(in % of GDP)
160
140
Private sector
Public sector
120
100
80
60
Monetary
expansion
Lender of last
resort
External deficits financed
by inflow of first FDIs than
of portfolio investments
None (IMF to some extent)
40
20
0
Turkey
48
Brasil
India
49. IV.2 Lessons can be learned: regulation and
policy
Regulation and policy institutions can help to avoid crises…
e.g. by fulfilling the role of Lender of Last Resort
Regulation to make credit growth less pro-cyclical (e.g. Basel III: counter cyclical
capital buffers)
Expectation shift by creation of OMTs by the ECB
- The ECB promises to do ‘whatever it takes’ (i.e. buy potentially unlimited amounts of
sovereign bonds) to prevent a forced EMU exit of member states
49
50. IV.2 Lessons to be learned: regulation and
policy
… or cause them
Example 1: “regulation” creating destructive incentives:
- Role of rating agencies partly based on regulatory definition of risk weighted assets
- Stress test in financial sector leading to further deleveraging
Example 2: the Tulip mania in the Netherlands in 1620s
- Change in legislation with respect to tulip futures and options contracts
Example 3: financial deregulation after the ‘80s
- E.g. originate and distribute model via financial engineering (packaging and selling risk)
- Volcker: ‘The only useful financial innovation in the past 30 years was the Automated
Teller Machine (ATM)’
50
51. IV.3 Lessons for quantitative risk
management: the illusion of safety
Black Swans: an event with a digital probability distribution is virtually
unmanageable
“Fat tail” risks can be addressed by using appropriate alternative
distributions
However, statistical distributions are not stable (invariant) over time
51
Are “fragile” under stress (Taleb Nassim)
Correlations in crisis times tend to rise
What was thought to be unlikely, is not unlikely at all
Probability distributions tend to change just at the time they are needed
52. IV.3 The example of Value at Risk
Consider a portfolio consisting of assets A and B with equal weights.
The variances and covariance of their returns are respectively Var(A), Var(B)
and Covar(A,B)
The portfolio return variance then equals
Var (portfolio) =
𝑉𝑎𝑟 𝐴 +𝑉𝑎𝑟
𝐵 +2 𝐶𝑜𝑣𝑎𝑟(𝐴,𝐵)
2
This means that the variance (and hence the standard deviation) of the
portfolio return increases as the correlation between the assets increases,
all else equal.
52 This is precisely what happens in financial crises
53. IV.3 The example of Value at Risk
Equity volatility increases sharply in times of financial crises
Implied volatility spikes in times of crises…
…such as the fall of Lehman (2008)
(in %)
(implied volatility in %)
90
90
80
80
70
70
60
60
50
50
40
40
30
30
20
20
10
10
Eurostoxx 50
Dax
Eurostoxx 50
S&P500
0
0
53
Dax
S&P500
54. IV.3 Value at Risk of 24.9%...
A (normal) return distribution with mean 8% and standard
deviation of 20%
0.02
0.015
0.01
0.005
0
-60
-50
-40
-30
-20
-10
0
10
20
30
Value at Risk = 24.9% (with 5% probability)
54
40
50
60
55. … is really 41.3% in times of crisis
(Normal) return distributions with mean 8%
0.02
STDEV = 30%
STDEV = 20%
0.015
0.01
0.005
0
-60
-50
-40
-30
-20
-10
0
10
Value at Risk = 41.3% (with 5% probability)
55
20
30
40
50
60
56. IV.3 Lesson for quantitative risk modelling
Time dependency of correlation data creates an illusion of safety
“In complex systems, such as financial systems, correlations are not
constant but vary in time. [...] The average correlation among stocks
scales linearly with market stress. [...] Consequently, the diversification
effect which should protect a portfolio melts away in times of market loss,
just when it would most urgently be needed.” (Preis et al. (2012))
One way to address time-dependency of risk models could be using statedependent correlation data, i.e. conditional (state dependent) instead of
unconditional correlations
56
57. Overview
Part I: Crises are more than ‘volatility’
Part II: Crises are inevitable
Part III: History repeats itself
Part IV: Lessons can be learned
Part V: Takeaways
57
58. V Takeaways
Market movements during financial crises are much stronger than normal volatility
Occasional financial crises/bubbles are unavoidable
Our financial system is inherently unstable (debt money, fractional reserve
banking,…)
There are limits to rational behaviour of economic agents
Credit cycles are at the core of most financial crises
A typical crises consists of several phases: displacement, boom, euphoria and bust
This model can be applied to identify potential new crises
Financial regulation can mitigate crises, or exacerbate them
58 A false sense of safety in quantitative risk management should be avoided