Information Gathering in Government Bailout Decision: An Experiment Ayung Tseng December 8, 2009 1.

Information Gathering in Government Bailout Decision: An Experiment

Ayung Tseng

December 8, 2009

1

Motivation

• Senator Jim Bunning “…you have decided that just about every large bank, investment bank, insurance company, and even some industrial companies are too big to fail. Rather than making management, shareholders, and debt holders feel the consequences of their risk-taking, you bailed them out. In short, you are the definition of moral hazard.”

• Senator Judd Gregg “If you hadn't been there, and hadn't been willing to take extraordinary action last fall, last winter, and even early spring … it's very likely we would be experiencing a depression, or certainly a recession that would be radically more severe than what we experienced…”

• The Federal Reserve held an emergency meeting during the weekend of Sept. 13-14, 2008, with top Washington policymakers and executives of major financial institutions to discuss the future of Lehman Brothers. Then on Sept 15, Lehmann filed bankruptcy.

2

Research Question

• What information gathering regime will enable government to receive the most complete and accurate information within a short period of time?

• This paper uses a laboratory experiment to explore ways of gathering information in government bailout decision during times of crises, especially focusing on the recent financial crisis in the banking industry.

• The main treatment in this experiment is differentiated by ways of information gathering: mandatory, voluntary, and communicative disclosure.

3

Asset Market Realized return RX

(exogenous)

Money MarketInterest rate Rf

(exogenous)

Consumers Banks

GovernmentMoney Supply

Investment

Set Rf rate

Lending at Rf Deposit at Rf

Consumer LoanRf + 5%

Leveraged 250%

Leveraged 180%

10% Deposit 90% Asset

50% Deposit 50% Asset

100% DepositRisk

Averse

Capital Ratio

Initial endowment(exogenous)

Participation in Interbank Market

Interbank Market

Rf

Investment

Financing

Leveraged Size

Deposits and Loans

distributed by leveraged size

Simulation of Financial Crisis

T1 Base yearRf = 3.0%

T2RX = (-50%)Rf = 0.1%

T3RX = (-20%)Rf = 0.1%

T4RX = 10%Rf = 0.1%

• Asset market value $2,032

• Money Supply $1,436

• Consumer market value $950

• Consumer avg. leverage 1.9

• Banking industry value $1,925

• Banking avg. leverage 19.3

• Asset market value $1,016

• Money Supply $749

• Consumer market total assets $363

• Consumers face margin calls and liquidate assets

• Banking industry value $1,139

• Some banks claim bankrupt and default on debts and deposits

• Asset market value $812

• Money Supply $384

• Consumer market total assets $23

• Consumers default on loans and claim bankrupt

• Banking industry value $641

• More banks claim bankrupt and default on debts and deposits

• Asset market value $894

• Money Supply $445

• Consumer market total assets $96

• Consumer market starts recovery

• Banking industry value $631

• Banking industry is still affected from interbank debt defaults

Simulation Results

The simulation shows the consequences of no bailout action and identifies the characteristics of the optimal bailout decision.

Experimental Design

Information structure:

•Aggregate economy loss: Public information

•Pre-crisis market composition: Public information

•Post-crisis loss distribution: Private information

Main treatments:

•Period 1-2: Mandatory disclosure without reputation effects

•Period 3-4: Voluntary then mandatory disclosure

•Period 5-6: Voluntary then communicative disclosure

•Period 7-8: Mandatory disclosure with reputation effects

6

Experimental Design (Cont.)

Conflicts of Interests:

• Government: minimize the social impact of financial crisis

• Banks: maximize shareholders’ value

Government bailout decision rule:

• If the sum of total loss claims submitted by all banks exceeds 120% of the economy’s total loss, no bailout funds are distributed.

• If the sum of total loss claims is less than or equal to 120% of the economy’s total loss, the government will allocate funds to the bank that reports the largest loss amount first, followed by the second largest, and so on until the bailout fund is exhausted.

7

Experiment Results

• Mandatory disclosure without reputation effects has the highest chance achieving the optimal decision.

• In the voluntary disclosure regime, banks’ revised disclosure amount is usually higher than its voluntary disclosure amount. A Pareto-suboptimal equilibrium leads to no bailout funds distributed.

• The communicative or collusive disclosure achieved an equal distribution of resources, deviated from the government goal of identifying the largest troubled bank.

• Mandatory disclosure with reputation effects is less effective than without reputation effects.

8

Country 1 Bailout Decision

9

98%

2%

25% 25%

68%

0%

20%

40%

60%

80%

100%

120%

0 1 2 3 4 5 6 7 8 9

Percentage of Actual Funds Allocation Matched with the Optimal Funds Allocation

Country 1

N N N

Country 1 Disclosure Amount

-100%-80%-60%-40%-20%

0%20%40%60%80%

100%120%140%

0 1 2 3 4 5 6 7 8 9 10 11 12

Total Claimed Loss Amount Difference from the Total Actual Loss Amount

Country 1

3-1 3-2 4-1 4-2 5-1 5-2 6-1 6-2 7 8

10

Country 2 Bailout Decision

98%

2%

74%

0%

20%

40%

60%

80%

100%

120%

0 1 2 3 4 5 6 7 8 9

Percentage of Actual Funds Allocation Matched with the Optimal Funds Allocation

Country 2

N N N N N

11

Country 2 Disclosure Amount

-100%-80%-60%-40%-20%

0%20%40%60%80%

100%120%140%

0 1 2 3 4 5 6 7 8 9 10 11 12

Total Claimed Loss Amount Difference from the Total Actual Loss Amount

Country 2

3-1 3-2 4-1 4-2 5-1 5-2 6-1 6-2 7 8

12

Country 3 Bailout Decision

98%

0% 2%0%

20%

40%

60%

80%

100%

120%

0 1 2 3 4 5 6 7 8 9

Percentage of Actual Funds Allocation Matched with the Optimal Funds Allocation

Country 3

N N N N N

13

Country 3 Disclosure Amount

-100%-80%-60%-40%-20%

0%20%40%60%80%

100%120%140%

0 1 2 3 4 5 6 7 8 9 10 11 12

Total Claimed Loss Amount Difference from the Total Actual Loss Amount

Country 3

3-1 3-2 4-1 4-2 5-1 5-2 6-1 6-2 7 8

14

Country 4 Bailout Decision

98%

0%

82%

25% 22%

73%

0%

20%

40%

60%

80%

100%

120%

0 1 2 3 4 5 6 7 8 9

Percentage of Actual Funds Allocation Matched with the Optimal Funds Allocation

Country 4

N N

15

Country 4 Disclosure Amount

-100%-80%-60%-40%-20%

0%20%40%60%80%

100%120%140%

0 1 2 3 4 5 6 7 8 9 10 11 12

Total Claimed Loss Amount Difference from the Total Actual Loss Amount

Country 4

3-1 3-2 4-1 4-2 5-1 5-2 6-1 6-2 7 8

16

Conclusion and Improvement

• During a crisis, moral hazard becomes an issue in information gathering process. This experiment found that the most effective and efficient way to gather information in such a time-constrained situation was mandatory anonymous disclosure. Based on this experiment result, the mandatory anonymous disclosure has a 50% chance leading to the optimal bailout decision.

• Potential improvement: 1) control six participants in each country, 2) give bank’s endowment in dollar amount, 3) revise the bankruptcy protection rule, 4) provide the probability of bankruptcy, 5) run this experiment on computer, 6) design other more effective information gathering regimes.

• Any ideas or suggestions?

17