Wei Xiong, Princeton University 5th Annual CARISMA Conference London, February 3, 2010 Post-Crisis Perspectives: Bubbles, Financial Institutions, and Asset.

Wei Xiong, Princeton University

5th Annual CARISMA Conference

London, February 3, 2010

Post-Crisis Perspectives: Bubbles, Financial Institutions, and Asset

Markets

Wei Xiong 2

Important Lessons in Three Fronts

1. Understanding asset bubbles is vital for understanding the recent crisis.

– Intricate interaction between the housing bubble and credit market.

2. Financial institutions are vulnerable to bubbles.– Various forms of externalities when the housing bubble went

bust.

3. Financial institutions have profound impacts on asset markets and asset prices.

– Unprecedented volatility and illiquidity across asset markets when financial institutions ran into crisis.

Wei Xiong 3

Part I: Bubbles as the Seed of the Crisis

• The crisis followed two great bubbles: the Internet bubble and the housing bubble.– The burst of the Internet bubble led to an expansionary US monetary

policy for a prolonged period.– The low interest rate together with financial innovations such as

securitization of subprime mortgages had fueled housing speculation and led to a great housing bubble.

– The burst of the housing bubble eroded the balance sheets of many financial institutions, and eventually dragged down the world economy.

Wei Xiong 4

The Housing Bubble

Mar-75 Jun-77 Sep-79 Dec-81 Mar-84 Jun-86 Sep-88 Dec-90 Mar-93 Jun-95 Sep-97 Dec-99 Mar-02 Jun-04 Sep-06 Dec-080

50

100

150

200

250

300

350

Real Home Price Indices

US

Arizona

California

Nevada

Florida

Wei Xiong 5

What Do We Know About Bubbles?

• It is common to see price bubbles in new technologies, such as railroad, radio, biotech, Internet, securitization, …

• New concepts stimulate dispersion of opinions and therefore speculation between optimists and pessimists.

• Miller (1977)– When investors have heterogeneous beliefs about asset fundamentals

and are constrained from short-selling, optimists determine price and drive overvaluation.

– Supported by evidence that measures of belief dispersion tend to predict lower stock returns, e.g., Diether, Malloy, and Scherbina (2002), and Chen, Hong and Stein (2002).

• Harrison and Kreps (1978) – In a dynamic setting with fluctuating beliefs, asset prices could be higher

than the current optimists’ valuation, because of the option to resell to even more optimistic buyers in the future.

– The value of the resale option represents a speculative component.

Wei Xiong 6

Overconfidence and Speculative Bubbles (Scheinkman and Xiong, 2003)

• Overconfidence (or inefficiencies in learning) provides a reasonable way to parameterize heterogeneous beliefs.– Belief dispersion persists and fluctuates with information flow.

• Volatility of investors’ disagreement drives both asset trading and bubble component in price, thus correlating the two.– Capturing the widely observed phenomenon that historical bubbles

were all associated with trading frenzies, e.g., Ofek and Richardson (2003), Lamont and Thaler (2003), and Cochrane (2003).

– Consistent with the large number of investment-home purchases in 2004-2006 in FL, NV, and AZ.

• Transaction cost has a second-order effect on price bubbles.– Bubbles could emerge in housing markets despite high trading cost.– Tobin’s tax may not work as intended in curbing bubbles.

Wei Xiong 7

Credit Booms and the Housing Bubble

• Credit expansion, e.g., securitization of innovative subprime mortgages, played an important role in fueling the housing bubble.– Mian and Sufi (2009)– Consistent with the view of Kindleberger (1978).

• A wave of effort to incorporate effects of credit in asset prices – Geanakoplos (2009) and Brunnermeier and Pedersen (2009)

• When optimists need to borrow from pessimists, belief divergence can cause credit to tighten and therefore asset prices to fall. – Simsek (2009):

• How did optimistic housing speculators manage to get credit from (not so optimistic) creditors during the housing bubble?

IMPO

RTAN

T

IMPO

RTAN

T

Wei Xiong 8

Financing Bubbles (He and Xiong, 2009)

• Consider an asset (house) with a two-period fundamental tree:– Optimists believe the up probability is π;– Pessimists believe the probability is ρ< π.

• Long-term collateralized debt allows optimists totrade promises across the three final states for aninitial loan from pessimists.

• Short-term debt, via contingent refinancing, makes it possible to trade promises across all four possible paths.

– The increased contract flexibility allows optimists to take a higher leverage

• Higher belief dispersion leads to a higher leverage and a shorter debt maturity by optimists, together with a higher asset price.

• Examples:– Short-term hybrid mortgages used by over 75% of subprime borrowers in 2003-2007.– Repos and CPs used by financial institutions.

u2f

uf

f udf

df

d2f

Wei Xiong 9

US Primary Dealer Mean Leverage

• Source: Adrian and Shin (2009)

Wei Xiong 10

Repos, Financial CP, and M2O vernight repos , F inanc ia l C P and M 2

(weekly, J u ly 6 1994 as base da te )

A pr 29 2009

M ar 19 2008

A ug 8 2007

2 .37

0 .0

1 .0

2 .0

3 .0

4 .0

5 .0

6 .0

7 .0

8 .0

Jul 6 1994

Jan 18 1995

Aug 2 1995

Feb 14 1996

Aug 28 1996

Mar 12 1997

Sep 24 1997

Apr 8 1998

Oct 21 1998

May 5 1999

Nov 17 1999

Jun 7 2000

Dec 20 2000

Jul 4 2001

Jan 16 2002

Jul 31 2002

Feb 12 2003

Aug 27 2003

Mar 10 2004

Sep 22 2004

Apr 6 2005

Oct 19 2005

May 3 2006

Nov 15 2006

May 30 2007

Dec 12 2007

Jun 25 2008

Jan 7 2009

O vernight repo

F inanc ia l C P

M 2

Wei Xiong 11

How to Detect the Next Bubble?

• Central banks and regulators have increasingly recognized the danger of asset bubbles to financial markets.

• But thirty-years of academic research only finds weak evidence of predictability in financial prices, i.e., it is hard to predict bubbles!• Even ex post identification of bubbles is often controversial because of the

difficulty in measuring asset fundamentals. • Debates on each historical bubble.

• It is useful to extract information from dimensions beyond prices.• Trading frenzy; • Heavy participation of inexperienced investors;• Large leverage taken by both investors and institutions; • Short-term credit booms;• And, more …

IMPO

RTAN

T

IMPO

RTAN

T

Wei Xiong 12

The Chinese Warrants Bubble (Xiong and Yu, 2008)

• China introduced a new warrants market in 2005-2008. • The life of Wuliang put warrant (strike price 8 Yuan):

0

10

20

30

40

50

60

70

80

0

1

2

3

4

5

6

7

8

9

Ap

r-06

Jun-0

6

Au

g-06

Oc

t-06

De

c-06

Fe

b-07

Ap

r-07

Jun-0

7

Au

g-07

Oc

t-07

De

c-07

Fe

b-08

Sto

ck p

rice

(Y

ua

n)

Pu

t p

rice

(Y

ua

n)

Put price Stock price

Wei Xiong 13

The Chinese Warrants Bubble (Xiong and Yu, 2008)

• Identifying the fundamental of Wuliang put warrant:

0

1

2

3

4

5

6

7

8

9

Ap

r-06

May

-06

Jun-

06

Jul-0

6

Au

g-06

Sep-

06

Oct-

06

No

v-06

De

c-06

Jan-

07

Feb-

07

Mar

-07

Ap

r-07

May

-07

Jun-

07

Jul-0

7

Au

g-07

Sep-

07

Oct-

07

No

v-07

De

c-07

Jan-

08

Feb-

08

Mar

-08

Pri

ce (Y

uan

)

Put price Strike price Fundamental upper bound Black-Scholes price

Wei Xiong 14

The Chinese Warrants Bubble (Xiong and Yu, 2008)

• In a sample of 16 put warrants, the price bubbles are associated with trading frenzies.

Wei Xiong 15

The Last Day of WanHua Put Warrant

Wei Xiong 16

Lessons from the Chinese Warrants Bubble

• The non-zero price traded at the very last minute of each put warrant indicates the existence of naïve investors in the market.

• The clear downward trend in prices indicates smart investors who recognize that as maturity approaches, there is less time to resell their warrants.

• The interaction between the naïve and smart investors ultimately drive the spectacular warrants bubble.

Wei Xiong 17

Speculative Behavior in the US Housing Bubble

20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 200.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1Share of Investment Property

National

Arizona

California

Florida

Nevada

Wei Xiong 18

Speculative Behavior in the US Housing Bubble

200501

200503

200505

200507

200509

200511

200601

200603

200605

200607

200609

200611

200701

200703

200705

200707

200709

200711

200801

200803

200805

200807

200809

200811

200901

200903

200905

2009070

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

Share of Second Homes

NationalArizonaCaliforniaFloridaNevada

Wei Xiong 19

Speculative Behavior in the US Housing Bubble

200501

200503

200505

200507

200509

200511

200601

200603

200605

200607

200609

200611

200701

200703

200705

200707

200709

200711

200801

200803

200805

200807

200809

200811

200901

200903

200905

2009070

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

Foreclosure Rate of Investment Property

USAZCAFLNV

Wei Xiong 20

Speculative Behavior in the US Housing Bubble

200501

200503

200505

200507

200509

200511

200601

200603

200605

200607

200609

200611

200701

200703

200705

200707

200709

200711

200801

200803

200805

200807

200809

200811

200901

200903

200905

2009070

0.02

0.04

0.06

0.08

0.1

0.12

Foreclosure Rate of Second Homes

USAZCAFLNV

Wei Xiong 21

Part II: The Fragility of Financial Institutions

• The shadow banking system, e.g., investment banks, SIVs, hedge funds, grew rapidly in the last decade. – Financial innovations, such as repos and other short-term debt

contracts, allowed many optimistic institutions to obtain large amount of credit from the rest of the economy.

• The burst of the housing bubble eroded the balance sheets of these institutions.– In particular, their asset-liability maturity mismatches expose

them to runs and other externalities during the down turns.– Insights from corporate finance are very useful in understanding

financial institutions and asset markets.

Wei Xiong 22

Runs on Asset-Backed Commercial Paper

• Source: Federal Reserve Release

Wei Xiong 23

Dynamic Debt Runs (He and Xiong, 2009)

• A model of runs that integrates fundamental concerns and panics.• A firm invests in a long-term illiquid asset by using staggered short-term

debt.– Each creditor is locked in during his contract period, during which others’ contracts

could mature.– Debt run externality.

• Fear of a firm’s future rollover risk could lead to preemptive runs.– Trigger of rollover risk: deterioration of firm fundamental.– Coordination problem between creditors amplifies this concern.

• A unique debt run equilibrium– Each maturing creditor chooses to run at a high fundamental threshold.– The threshold depends on fundamental volatility, market liquidity, and debt

maturity.

• A tractable framework for dynamic coordination problems.– Features time-varying fundamental and a unique equilibrium.– No need for asymmetric information.

Wei Xiong 24

Debt Overhang on Financial Institutions

• When firms face difficulties in rolling over maturing debt, why don’t they issue equity instead?– Debt overhang and bankruptcy externality.

• He and Xiong (2009) on ``Rollover Risk and Credit Risk”– A firm can issue equity at market price to pay off rollover losses at any time.– The firm defaults when its equity price, which is endogenously determined by

firm fundamental and future rollover losses, drops to zero, a la Leland (1994).

• Deteriorating market liquidity exacerbates the conflict between debt and equity holders, causing earlier endogenous default by equity holders. – It not only increases liquidity premium, but also default premium.

• Short-term debt further amplifies the conflict.– A tradeoff in using short-term debt: cheaper financing cost and higher future

bankruptcy cost.

Wei Xiong 25

Other Externalities in the Financial System

• Liquidation externality– Morris and Shin (2004), Bernardo and Welch (2004), Brunnermeier

and Pedersen (2009)

• Leverage externality– Lorenzoni (2008)

• Network externality– Zawadowski (2009)

• The existence of these different forms of externalities in the financial system prompts macro-prudent risk management rather than over-reliance on micro-prudent risk management.– Brunnermeier, Crockett, Goodhart, and Shin (2009)– Acharya and Richardson (2009)

Wei Xiong 26

Part III: Impact of Financial Institutions on Asset Markets

• The standard finance theories attribute asset prices to rational expectations of assets’ future cash flow and systematic risk.

• The crisis again demonstrates the great influence of financial institutions on asset price dynamics.• Confirming the view advocated by Allen and Gale (2007) and

others.

Wei Xiong 27

The Boom and Bust of Commodities in 2008

Wei Xiong 28

Index Investing and the Financialization of Commodities (Tang and Xiong, 2009)

• Two popular views about the boom and bust of commodities:– A matter of supply & demand, e.g., Krugman (2008) and Hamilton (2008).– Financial speculation has distorted commodity prices, e.g., CFTC (2008) and

Masters (2008). • The debate ignores a fundamental financialization process of

commodities.– Prior to early 2000s, commodities were segmented from the broader

financial markets and from each other.– The rapid growth of index investment in commodities made commodities

behave much more like stocks, i.e., they are increasingly exposed to market-wide shocks and shocks to other commodities.

– As a result, the spillover effects of the recent financial crisis have contributed substantially to the large increase in commodity price volatility in 2008.

Wei Xiong 29

Return Correlation: Soybean-Oil

Wei Xiong 30

Return Correlation: Cotton-Oil

Wei Xiong 31

Return Correlation: Live Cattle-Oil

Wei Xiong 32

Return Correlation: Copper-Oil

Wei Xiong 33

Return Correlation between Soybean Complex and Oil

Wei Xiong 34

The Difference-in-Difference Analysis

• Significant differences in the exposures to shocks to stock markets, US dollar, and oil between individual non-energy commodities in the Goldman Sachs Commodity Index and Dow Jones Commodity Index and those off the indices.

Wei Xiong 35

Volatility of Commodities

Wei Xiong 36

Decomposing Oil VolatilityIM

PORTAN

T

IMPO

RTAN

T

Wei Xiong 37

Decomposing Non-energy Commodity Volatility

IMPO

RTAN

T

IMPO

RTAN

T

Wei Xiong 38

Many Important Questions Remaining

• Understanding bubbles:– Relationship between bubbles and credit markets

– How to predict bubbles?

• Understanding financial institutions– Externalities

– Compensation and incentives

– Macro-prudent risk management

• Impact of financial frictions on asset prices and the economy. – Volatility and risk premium

– Firm investment and economic growth