Competition policy as a tool for the macroprudential regulation of - - PowerPoint PPT Presentation

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

Competition policy as a tool for the macroprudential regulation of the banking sector

Massimo Molinari § (Joint with Edoardo Gaffeo)

§ University of Trento, Department of Economics

massimo.molinari@unitn.it Latsis Symposium 2012 - Economics on the Move - ETH, Zurich

September, 2012

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

(1) Motivation

◮ Can network structure be altered to improve network robustness? (Haldane, 2009);

(2) THEORETICAL PERSPECTIVE

◮ Macro-prudential regulation of the banking system (Hanson et al., 2009);

◮ time-varying and size-varying capital requirements, high(er)-quality capital,

dollars at capital (as opposed to capital ratios), contingent capital, a more tight regulation of debt maturity, the regulation of the shadow banking system

◮ Competition Policy (Vives, 2010);

◮ Trade-off between competition and stability

◮ Network models (Nier et al. (2007), Gai and Kapadia (2010).

◮ Non-linear (Inverted U-shape) relationship between connectivity and the

resilience of the system (3) CONTRIBUTION of the PAPER

◮ Examine the interaction between competition policy and macro-prudential

regulation using a network approach.

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Definition of a bank Building up the network

Bank_1 Bank_2 Bank_3 Bank_4 Bank_5 Bank_6 Bank_7 Bank_8 Bank_9 Bank_10 Bank_11 Bank_12 Bank_13 Bank_14 Bank_15 Bank_16 Bank_17 Bank_18 Bank_19 Bank_20 Bank_21 Bank_22 Bank_23 Bank_24 Bank_25

Figure: Homogenous Banking Network

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Definition of a bank Building up the network

The asset structure of bank i is made up as follows: Assets Liabilities Ai NWi Li Bi Di

Ai=External Assets, Li=Lending NWi=Networth, Bi=Borrowing, Di=Deposits Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Definition of a bank Building up the network

• Key-object in our agent-based laboratory: Flow matrix RF.
• Li = ∑n

j=1 RFij (horizontal summation) where Li is the total lending

• f bank i
• Bi = ∑n

j=1 RF T ij (vertical summation) where Bj is the total

borrowing of bank j

• Once we have retrieved from RF Bi and Li ∀ i, we built each bank

asset structure in the following way:

◮ Ai = αLi ◮ NWi = β[Ai + Li] ◮ Di = Ai + Li − NWi − Bi Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Shock and propagation I Loss cascade Characterization of the network

• We now introduce a shock Si = γAi that wipes out some or all of the

external assets of bank i and we let the system adjust to it.

◮ Whenever Ai drops, NWi is reduced by the same amount. Three

scenarios are possible:

◮ If NWi − Si > 0 then the bank survives and the shock is

fully absorbed by the first bank.

◮ If NWi − Si = 0, the bank fails but no other bank is affected

by it. All lenders and depositors get their money back.

◮ If NWi − Si < 0, bank i fails and losses are distributed

amongst creditor banks linked with bank i.

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Shock and propagation I Loss cascade Characterization of the network

◮ mechanism of shock transmission: suppose that the first bank defaults

and turns out to be unable to repay 70 percent of its loans. Each creditor will then lose 70 percent of the value of the loan made to that bank.

◮ Ai,R ↓=

⇒ NWi,R ↓(First-order Loss)

◮ Bi,R ↓ (Second-order loss)=

⇒ Lj,R+1 ↓= ⇒ NWj,R+1 ↓

◮ Di,R ↓ (Third-order loss)

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Shock and propagation I Loss cascade Characterization of the network

The network is fully characterized by the following set of parameters: Table: Description of Parameters

Parameters Description Benchmark Value Range of Variation n Number of Nodes (Banks) 25 p Probability of Connectivity 0.2 α External Assets to Interbank Lending Ratio 5 β Net-worth to Total Assets Ratio 0.01-0.07 γ Shock relative to External Assets of one bank 1

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

◮ central banks and antitrust authorities have the opportunity to design

the structure of the industry by choosing how banks are allowed to merge.

◮ Think about the case of Bankia in Spain ◮ The key-question is whether network structures can be modified to

strengthen the system resilience to shocks.

◮ merges change the topology of the sector changes for three reasons:

(1) larger banks are formed (2) the total number of active banks decrease (3) large banks are assumed to have more connections than small banks.

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

◮ We start at round 0 with a population of 25 homogeneous banks and

we simulate one merger at each of the following 9 rounds.

◮ three different M&A strategies, which we translate into three different

experimental treatments.

(1) T1: a merger is possible only between two small banks. (2) T2: at each stage one small bank is acquired by the same large bank. (3) T3: at each round the M&A process creates a new large bank, but the size of active large banks is bound to remain equal horizontally

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

Figure: Herfindahl Index

1 2 3 4 5 6 7 8 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 Merge−Round Herfindahl index Treatment I Treatment II Treatment III

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results Table: Summary table of the treatments

Rounds Ss Sl Ns Nl Ps Pl T1 T2 T3 T1 T2 T3 T1 T2 T3 T1 T2 T3 T1 T2 T3 T1 T2 T3 1 60 60 60 na na na 25 25 25 0.2 0.2 0.2 na na na 2 60 60 62.50 120 120 na 23 23 24 1 1 0.2 0.2 0.218 0.617 0.617 na 3 60 60 65.22 120 180 na 21 22 23 2 1 0.2 0.205 0.237 0.627 1 na 4 60 60 68.18 120 240 na 19 21 22 3 1 0.2 0.224 0.260 0.638 1 na 5 60 60 71.43 120 300 na 17 20 21 4 1 0.2 0.250 0.286 0.650 1 na 6 60 60 75 120 360 na 15 19 20 5 1 0.2 0.280 0.316 0.663 1 na 7 60 60 78.95 120 420 na 13 18 19 6 1 0.2 0.315 0.351 0.677 1 na 8 60 60 83.33 120 480 na 11 17 18 7 1 0.2 0.356 0.392 0.694 1 na 9 60 60 88.24 120 540 na 9 16 17 8 1 0.2 0.406 0.441 0.712 1 na Ss=Size of Small Banks, Ns=Number of Small Banks, Ps=Connectivity of Small Banks Sl=Size of Large Banks, Nl=Number of Large Banks, Pl=Connectivity of Large Banks NsSs + NlSl = 1500 ∀ Rounds and Treatments Average Number of Links=120 ∀ Rounds and Treatments

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

Deep Interbank Market and Undercapitalized System

1 2 3 4 5 6 7 8 9 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 Rounds Relative Asset Dislocation T3 T2 (Shock to a Large Bank) T1 (Shock to a Large Bank) 1 2 3 4 5 6 7 8 9 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 Rounds Relative Asset Dislocation T3 T2 (Shock to a Small Bank) T1 (Shock to a Small Bank)

Figure: β = 0.01, α = 2, Shock=40

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

Shallow Interbank Market and Undercapitalized System

1 2 3 4 5 6 7 8 9 1 1.5 2 2.5 Rounds Relative Asset Dislocation T3 T2 (Shock to a Large Bank) T1 (Shock to a Large Bank) 1 2 3 4 5 6 7 8 9 1 1.5 2 2.5 Rounds Relative Asset Dislocation T3 T2 (Shock to a Small Bank) T1 (Shock to a Small Bank)

Figure: β = 0.01, α = 5, Shock=50

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

Shallow Interbank Market and Capitalized System

1 2 3 4 5 6 7 8 9 1 1.5 2 2.5 Rounds Relative Asset Dislocation T3 T2 (Shock to a Large Bank) T1 (Shock to a Large Bank) 1 2 3 4 5 6 7 8 9 1 1.5 2 2.5 Rounds Relative Asset Dislocation T3 T2 (Shock to a Small Bank) T1 (Shock to a Small Bank)

Figure: β = 0.07, α = 5, Shock=50

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots Treatments Simulation Results

◮ Shock to small firms are more damaging for the system ◮ The lower β and the higher α, the more the three treatments produce

different results and hitting a large bank is less damaging for the system.

◮ This gap gradually narrows down as we increase the level of networth

in the system or shrink the interbank market. As a matter of fact, very little differences in treatments are detectable when the system is capitalized at 7

◮ Concentrated and yet Asymmetric Banking networks are better

equipped to deal with a shock

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

Treatment 2

2 4 6 8 10 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Rounds Relative Asset Dislocation Shock to a Large Bank−βl=0.01,βs=0.01 Shock to a Large Bank−βl=0.07,βs=0.01 2 4 6 8 10 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Rounds Relative Asset Dislocation Shock=50 Shock to a small Bank−βl=0.01,βs=0.01 Shock to a small Bank−βl=0.07,βs=0.01

Figure: Asset Dislocation with conditional capital requirements

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

Treatment 2

βopt

l

=

1 1+α 1 Pl(ns+nl−1)

βopt

s

=

1 1+α 1 Ps(ns+nl−1)

2 4 6 8 10 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 Rounds β βopt

s

βopt

l

2 4 6 8 10 10 20 30 40 50 60 Rounds Aggregate Equity NWopt−T2

s

NWopt−T2

l

NWopt−T2

agg

NW01−01

s

NW01−01

l

NW01−01

agg

Figure: Panel A: Optimal conditional capital requirements - Panel B: Aggregate and Partial Equity

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

Treatment 2

1 2 3 4 5 6 7 8 9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Rounds Relative Asset Dislocation Shock=50 Shock to a Large Bank−βl=βopt

l

,βs=βopt

s

1 2 3 4 5 6 7 8 9 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Rounds Defaults Defaultagg Defaults Defaultl

Figure: Asset Dislocation and Default Dynamics with optimal capital requirements- shock to a large bank

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

(1) CONCLUSIONS

◮ Topologies are not all alike: a concentrated and yet asymmetric

system is better geared to cope with an external shock.

◮ The extent of the damage to the system depends on the exposure

to interbank claims, the degree of connectivity, the structure of the network and capital requirements.⇒ capital requirements should be network-varying.

◮ Different shock-amplifying dynamics arise because flat capital

requirements force an inefficient allocation of net worth within the system.

◮ Once we introduce network-varying capital requirements, the

robustness of the system improves and this aligns the performance of different topologies.

◮ policy implication: the regulator shall closely monitor the

structure of the network and its evolution over time.

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

(2) EXTENSIONS and FUTURE WORK

◮ Which role for Large Banks? Net Borrowers or Net Lenders ◮ Multiple Shocks ◮ Scale-free Networks ◮ Liquidity Effects.

Gaffeo E. and Molinari M. Systemic Risk in Banking Networks

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

1 2 3 4 5 6 7 8 9 10 20 30 40 50 60 Rounds Aggregate Equity NWopt−T2

s

NWopt−T2

l

NWopt−T2

agg

NWopt−T1

s

NWopt−T1

l

NWopt−T1

agg

Figure: Aggregate networth with Optimal Capital Requirements-T1 vs T2

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

Treatment 1

1 2 3 4 5 6 7 8 9 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05 0.055 βopt

s

βopt

l

Figure: Optimal β

Motivations and Theory Background Banking Network Default dynamics Network structures and Competition Policy Conditional Capital Requirements Network-Varying Capital Requirements Final Remarks Additional Plots

Treatment 1

2 4 6 8 10 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Rounds Relative Asset Dislocation Shock=50 Shock to a Large Bank−βl=0.07,βs=0.01 2 4 6 8 10 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Rounds Relative Asset Dislocation Shock=50 Shock to a Large Bank−βl=βopt

l

,βs=βopt

s

Figure: Asset Dislocation - shock to a large bank