Imperfect Transparency and the Risk of Securitization

Transcription

1 Imperfect Transparency and the Risk of Securitization Seungjun Baek Florida State University June. 16, 2017

2 1. Introduction

3 Motivation Study benefit and risk of securitization

4 Motivation Study benefit and risk of securitization Benefit: relax borrowing constraints ( diversify idiosyncratic risk)

5 Motivation Study benefit and risk of securitization Benefit: relax borrowing constraints ( diversify idiosyncratic risk) Risk: solvency and funding risk

6 Motivation Study benefit and risk of securitization Benefit: relax borrowing constraints ( diversify idiosyncratic risk) Risk: solvency and funding risk Build a model of banking with securitization

7 Motivation Study benefit and risk of securitization Benefit: relax borrowing constraints ( diversify idiosyncratic risk) Risk: solvency and funding risk Build a model of banking with securitization Endogenize solvency and funding risk

8 Motivation Study benefit and risk of securitization Benefit: relax borrowing constraints ( diversify idiosyncratic risk) Risk: solvency and funding risk Build a model of banking with securitization Endogenize solvency and funding risk Securitization leads to socially excessive solvency and funding risk.

9 Key Innovations Endogenous Solvency and Funding Risk

10 Key Innovations Endogenous Solvency and Funding Risk Solvency Risk (Quality) : adverse selection between securitizers (originators) and projects (borrowers).

11 Key Innovations Endogenous Solvency and Funding Risk Solvency Risk (Quality) : adverse selection between securitizers (originators) and projects (borrowers). must exert costly effort.

12 Key Innovations Endogenous Solvency and Funding Risk Solvency Risk (Quality) : adverse selection between securitizers (originators) and projects (borrowers). must exert costly effort. Funding Risk : imperfect information about the quality (solvency) of assets.

13 Key Innovations Endogenous Solvency and Funding Risk Solvency Risk (Quality) : adverse selection between securitizers (originators) and projects (borrowers). must exert costly effort. Funding Risk : imperfect information about the quality (solvency) of assets. Greater transparency leads to a decrease in funding risk.

14 Key Innovations Securitization : create the pool of underlying projects (ABS), then trade ABSs

15 Key Innovations Securitization : create the pool of underlying projects (ABS), then trade ABSs = diversify idiosyncratic risk : the pledgeability of assets

16 Key Innovations Securitization : create the pool of underlying projects (ABS), then trade ABSs = diversify idiosyncratic risk : the pledgeability of assets

17 Preview of the Paper Key Objective : quality Profits = N.P.V leverage funding risk fire-sale price. }{{} losses from funding crises

18 Preview of the Paper Key Objective : quality Profits = N.P.V leverage funding risk fire-sale price. }{{} losses from funding crises Key Externalities : Externality 1: the benefit of transparency not internalized.

19 Preview of the Paper Key Objective : quality Profits = N.P.V leverage funding risk fire-sale price. }{{} losses from funding crises Key Externalities : Externality 1: the benefit of transparency not internalized. - funding risk depends on aggregate transparency of a portfolio (diversification via trade) - transparency also depends on choices of others.

20 Preview of the Paper Key Objective : quality Profits = N.P.V leverage funding risk fire-sale price. }{{} losses from funding crises Key Externalities : Externality 1: the benefit of transparency not internalized. - funding risk depends on aggregate transparency of a portfolio (diversification via trade) - transparency also depends on choices of others. Externality 2: the benefit of quality not internalized.

21 Preview of the Paper Key Objective : quality Profits = N.P.V leverage funding risk fire-sale price. }{{} losses from funding crises Key Externalities : Externality 1: the benefit of transparency not internalized. - funding risk depends on aggregate transparency of a portfolio (diversification via trade) - transparency also depends on choices of others. Externality 2: the benefit of quality not internalized. - the planner s objective = N.P.V. - higher leverage => high quality becomes less valuable.

22 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities.

23 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities. The composition of the asset side of the balace sheet matters : new dimension of macroprudential policy.

24 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities. The composition of the asset side of the balace sheet matters : new dimension of macroprudential policy. The recent financial crisis: lax lending standards Originate-to-distribute hypothsis : Diamond and Rajan (2009), Akerlof and Shiller (2010).

25 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities. The composition of the asset side of the balace sheet matters : new dimension of macroprudential policy. The recent financial crisis: lax lending standards Originate-to-distribute hypothsis : Diamond and Rajan (2009), Akerlof and Shiller (2010). Objections : Gorton (2008), Acharya, Schnabl, and Suarez (2013), Erel, Nadauld, and Stulz (2014).

26 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities. The composition of the asset side of the balace sheet matters : new dimension of macroprudential policy. The recent financial crisis: lax lending standards Originate-to-distribute hypothsis : Diamond and Rajan (2009), Akerlof and Shiller (2010). Objections : Gorton (2008), Acharya, Schnabl, and Suarez (2013), Erel, Nadauld, and Stulz (2014). Irrational optimism: Gennaioli, Shleifer and Vishny (2012,2013).

27 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities. The composition of the asset side of the balace sheet matters : new dimension of macroprudential policy. The recent financial crisis: lax lending standards Originate-to-distribute hypothsis : Diamond and Rajan (2009), Akerlof and Shiller (2010). Objections : Gorton (2008), Acharya, Schnabl, and Suarez (2013), Erel, Nadauld, and Stulz (2014). Irrational optimism: Gennaioli, Shleifer and Vishny (2012,2013). In my model, securitizers are rational, and do not have information advantage over other investors.

28 Related Literature/Key Contributions Financial stability and systemic risk. Lorenzoni (2008), Stein (2012) : need to limit the use of short-term liabilities. The composition of the asset side of the balace sheet matters : new dimension of macroprudential policy. The recent financial crisis: lax lending standards Originate-to-distribute hypothsis : Diamond and Rajan (2009), Akerlof and Shiller (2010). Objections : Gorton (2008), Acharya, Schnabl, and Suarez (2013), Erel, Nadauld, and Stulz (2014). Irrational optimism: Gennaioli, Shleifer and Vishny (2012,2013). In my model, securitizers are rational, and do not have information advantage over other investors. Transparency versus Opaqueness Dang, Gorton, and Holmstrom (2009, 2010, 2012)

29 1. State of Nature 2. Securitizers 3. Inside Investors 4. Outside Investors Model

30 Overview of the Model 3 dates, and initial wealth of securitizers, inside investors, and outside investors: 1, Y S, Y I respectively. State of Nature Technology Preference

31 Figure: State of Nature (Securitization)

32 Figure: Preference, Technology (Securitization)

33 Securitizer Can choose borrowing d up to the borrowing constraint d q

34 Definition (Securitization) Pool assets across islands to create ABS.

35 Definition (Securitization) Pool assets across islands to create ABS. Trade: fully diversifies idiosyncratic risk in the economy.

36 Definition (Securitization) Pool assets across islands to create ABS. Trade: fully diversifies idiosyncratic risk in the economy.

37 Definition (Securitization) Pool assets across islands to create ABS. Trade: fully diversifies idiosyncratic risk in the economy. Create safe cash flows -> increases pledgeability

38 Definition (Securitization) Pool assets across islands to create ABS. Trade: fully diversifies idiosyncratic risk in the economy. Create safe cash flows -> increases pledgeability Source of externality: diversifies the benefit of transparency

39 Definition (Securitization) Pool assets across islands to create ABS. Trade: fully diversifies idiosyncratic risk in the economy. Create safe cash flows -> increases pledgeability Source of externality: diversifies the benefit of transparency Transparency of their portfolio depends on aggregate transparency

40 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0.

41 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0. In s = g, it returns µ (S) > 0.

42 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0. In s = g, it returns µ (S) > 0. Quality Choice : long-term solvency risk λ.

43 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0. In s = g, it returns µ (S) > 0. Quality Choice : long-term solvency risk λ. Quality : 1 λ

44 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0. In s = g, it returns µ (S) > 0. Quality Choice : long-term solvency risk λ. Quality : 1 λ Cost of quality: ι 1 λ, where ι > 0.

45 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0. In s = g, it returns µ (S) > 0. Quality Choice : long-term solvency risk λ. Quality : 1 λ Cost of quality: ι 1 λ, where ι > 0. Transparency Choice : The accuracy of signal r i about the individual state s i : f (r i = s i s i ) = α, f (r i = s i s i ) = 1 α.

46 Securitizer : Quality and Transparency choice once and for all at t=1 An Individual underlying project defaults in s = b, yielding 0. In s = g, it returns µ (S) > 0. Quality Choice : long-term solvency risk λ. Quality : 1 λ Cost of quality: ι 1 λ, where ι > 0. Transparency Choice : The accuracy of signal r i about the individual state s i : f (r i = s i s i ) = α, f (r i = s i s i ) = 1 α. Cost of transparency: ι α, where ι α > 0.

47 Inside Investors (Short-term lenders to securitizers) Pursue safety assets.

48 Inside Investors (Short-term lenders to securitizers) Pursue safety assets. Only safe cash flows q are pledgeable to IIs. (no default at t = 2 ).

49 Inside Investors At date 1, IIs lend up to q.

50 Inside Investors At date 1, IIs lend up to q. At date 2,

51 Inside Investors At date 1, IIs lend up to q. At date 2, Receive signal φ (accuracy of which α = α j dj). => Form a posterior about the aggregate state.

52 Inside Investors At date 1, IIs lend up to q. At date 2, Receive signal φ (accuracy of which α = α j dj). => Form a posterior about the aggregate state. Decide whether to pay χ to learn about the aggregate state. (choose the early resolution of uncertainty)

53 Inside Investors At date 1, IIs lend up to q. At date 2, Receive signal φ (accuracy of which α = α j dj). => Form a posterior about the aggregate state. Decide whether to pay χ to learn about the aggregate state. (choose the early resolution of uncertainty) Decide whether to keep funding (receive R at t = 3) or not (receive 1 at t = 2).

54 Outside Investors at t=2

55 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets.

56 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets. The first order condition: q i r i di }{{} A = υ (M), (1) q gains from new projects }{{} gains from buying fire-sold assets

57 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets. The first order condition: q i r i di }{{} A = υ (M), (1) q gains from new projects }{{} gains from buying fire-sold assets

58 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets. The first order condition: q i r i di }{{} A = υ (M), (1) q gains from new projects }{{} gains from buying fire-sold assets q = q r i di. We may also introduce unsophisticated OIs who can t distinguish the good signal from the bad signal. q = q g = q b

59 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets. The first order condition: q i r i di }{{} A = υ (M), (1) q gains from new projects }{{} gains from buying fire-sold assets q = q r i di. We may also introduce unsophisticated OIs who can t distinguish the good signal from the bad signal. q = q g = q b Play a greater role when fire sales M is greater.

60 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets. The first order condition: q i r i di }{{} A = υ (M), (1) q gains from new projects }{{} gains from buying fire-sold assets q = q r i di. We may also introduce unsophisticated OIs who can t distinguish the good signal from the bad signal. q = q g = q b Play a greater role when fire sales M is greater. Analogous to Shleifer and Vishny (1992)

61 Outside Investors at t=2 Observe the distribution of signal {r i } about the quality of underlying assets. The first order condition: q i r i di }{{} A = υ (M), (1) q gains from new projects }{{} gains from buying fire-sold assets q = q r i di. We may also introduce unsophisticated OIs who can t distinguish the good signal from the bad signal. q = q g = q b Play a greater role when fire sales M is greater. Analogous to Shleifer and Vishny (1992) f = q g q b, and f M < 0.

62 Equilibrium

63 Equilibrium : Inside Investors (Lenders to Securitizers) t = 2, roll-over decisions : Lemma (Learning and Funding) (i) In S = M, IIs keep funding if and only if they learn about S.

64 Equilibrium : Inside Investors (Lenders to Securitizers) t = 2, roll-over decisions : Lemma (Learning and Funding) (i) In S = M, IIs keep funding if and only if they learn about S.

65 Equilibrium : Inside Investors (Lenders to Securitizers) t = 2, roll-over decisions : Lemma (Learning and Funding) (i) In S = M, IIs keep funding if and only if they learn about S. Intuition: for risk-averse IIs, E [U(C S 3 ) φ] }{{} expected utility under uncertainty < U (dn) }{{} certain utiliy with the cash withdrawal

66 Equilibrium : Inside Investors (Lenders to Securitizers) t = 2, roll-over decisions : Lemma (Learning and Funding) (i) In S = M, IIs keep funding if and only if they learn about S. Intuition: for risk-averse IIs, E [U(C S 3 ) φ] }{{} expected utility under uncertainty < U (dn) }{{} certain utiliy with the cash withdrawal With the early resolution of uncertainty in S = M, U(C M 3 ) > U (dn).

67 Equilibrium : Inside Investors (Lenders to Securitizers) t = 2, roll-over decisions : Lemma (Learning and Funding) (i) In S = M, IIs keep funding if and only if they learn about S. Intuition: for risk-averse IIs, E [U(C S 3 ) φ] }{{} expected utility under uncertainty < U (dn) }{{} certain utiliy with the cash withdrawal With the early resolution of uncertainty in S = M, U(C M 3 ) > U (dn). (ii) In S = L, all IIs stop funding.

68 Equilibrium : Inside Investors Lemma (i) Let P denote the fraction of IIs who learn about the state with the costly extra information in S = M. Then P α > 0;

69 Equilibrium : Inside Investors Lemma (i) Let P denote the fraction of IIs who learn about the state with the costly extra information in S = M. Then P α > 0; (ii) Let Λ denote funding risk, which is defined as the fraction of IIs quit funding. Λ M = (1 P(α)) }{{}, IIs who do not learn stop funding Λ L = 1.

70 Equilibrium : Inside Investors Lemma (i) Let P denote the fraction of IIs who learn about the state with the costly extra information in S = M. Then P α > 0; (ii) Let Λ denote funding risk, which is defined as the fraction of IIs quit funding. Λ M = (1 P(α)) }{{}, IIs who do not learn stop funding Λ L = 1.

71 Equilibrium : Inside Investors Lemma (i) Let P denote the fraction of IIs who learn about the state with the costly extra information in S = M. Then P α > 0; (ii) Let Λ denote funding risk, which is defined as the fraction of IIs quit funding. Λ M = (1 P(α)) }{{}, IIs who do not learn stop funding Λ L = 1. Intuition: In S = L, learning is less likely with greater transparency. If signal φ is transparent, paying for the additional information does not add much value.

72 Equilibrium : Inside Investors Lemma (i) Let P denote the fraction of IIs who learn about the state with the costly extra information in S = M. Then P α > 0; (ii) Let Λ denote funding risk, which is defined as the fraction of IIs quit funding. Λ M = (1 P(α)) }{{}, IIs who do not learn stop funding Λ L = 1. Intuition: In S = L, learning is less likely with greater transparency. If signal φ is transparent, paying for the additional information does not add much value. In S = M, learning is more likely.

73 Equilibrium : Inside Investors Lemma (i) Let P denote the fraction of IIs who learn about the state with the costly extra information in S = M. Then P α > 0; (ii) Let Λ denote funding risk, which is defined as the fraction of IIs quit funding. Λ M = (1 P(α)) }{{}, IIs who do not learn stop funding Λ L = 1. Intuition: In S = L, learning is less likely with greater transparency. If signal φ is transparent, paying for the additional information does not add much value. In S = M, learning is more likely. transparency increases the private opportunity costs of the cash withdrawal, leading IIs to learn and keep funding.

74 Equilibrium : Securitizer (Fire-sales at t=2) If some IIs stop funding, asset liquidation occurs in a resale market, } Λ {{ d } short-term liabilities = κ q }{{} proceeds from fire-sales, where κ is the fraction liquidated κ = Λd q.

75 Equilibrium : Securitizer (Fire-sales at t=2) If some IIs stop funding, asset liquidation occurs in a resale market, } Λ {{ d } short-term liabilities = κ q }{{} proceeds from fire-sales, where κ is the fraction liquidated κ = Λd q. (Collateral constraint) Since max[λ] 1 and max[κ] 1, d min[q] = q.

76 Equilibrium : Securitizer j (t=1) At date 1:

77 Equilibrium : Securitizer j (t=1) At date 1: choose borrowing, d, quality λ, and transparency α for each project.

78 Equilibrium : Securitizer j (t=1) At date 1: choose borrowing, d, quality λ, and transparency α for each project. pooling and trading ABS

79 Taking underlying asset price q and α as given, each chooses borrowing d, quality λ and transparency α to maximize: [E (µ) Rd ι λ (λ) ι α (α)] }{{} E [L(λ, α)]d }{{}, n.p.v. of projects net losses from fire-sales subject to the collateral constraint,

80 Taking underlying asset price q and α as given, each chooses borrowing d, quality λ and transparency α to maximize: [E (µ) Rd ι λ (λ) ι α (α)] }{{} E [L(λ, α)]d }{{}, n.p.v. of projects net losses from fire-sales subject to the collateral constraint, E [µ] =pµ + (1 p)(ηµ B + (1 η)(1 λ) µ B ), }{{} quality

81 Taking underlying asset price q and α as given, each chooses borrowing d, quality λ and transparency α to maximize: [E (µ) Rd ι λ (λ) ι α (α)] }{{} E [L(λ, α)]d }{{}, n.p.v. of projects net losses from fire-sales subject to the collateral constraint, E [µ] =pµ + (1 p)(ηµ B + (1 η)(1 λ) µ B ), }{{} quality L(λ, α) }{{} net losses from fire-sales = }{{} Λ { funding risk µ(λ) q R }{{} opportunity costs of liquidating a unit of ABS }.

82 Equilibrium ABS price depends on the distribution of signal {r i } q =((1 λ) α + λ (1 α)) q g + ((1 λ) (1 α) + λα) q }{{}}{{} weight on q g weight on q b b.

83 Equilibrium ABS price depends on the distribution of signal {r i } q =((1 λ) α + λ (1 α)) q g + ((1 λ) (1 α) + λα) q }{{}}{{} weight on q g weight on q b When transparency α is higher, q is closely correlated with its quality q (1 λ) (2α 1) f, where f = q g q b. b.

84 Equilibrium ABS price depends on the distribution of signal {r i } q =((1 λ) α + λ (1 α)) q g + ((1 λ) (1 α) + λα) q }{{}}{{} weight on q g weight on q b When transparency α is higher, q is closely correlated with its quality q (1 λ) (2α 1) f, where f = q g q b. Transparency raises q greater when quality is better q α (1 2λ) f b.

85 Constrained Effi ciency Constrained effi ciency: no better information than the private sector. debt must be safe. The social planner s problem: max W = U S + U II + U OI = E [µ λ] Rd ι λ (λ) ι α (α) E [υ(λd) + (A R) (Λd)] }{{} real losses from fire sales subject to the same collateral constraint.

86 Constrained Effi ciency Constrained effi ciency: no better information than the private sector. debt must be safe. The social planner s problem: max W = U S + U II + U OI = E [µ λ] Rd ι λ (λ) ι α (α) E [υ(λd) + (A R) (Λd)] }{{} real losses from fire sales subject to the same collateral constraint. υ(λd) : fire-sold assets lose value (Shleifer and Vishny, 1992)

87 Constrained Effi ciency Constrained effi ciency: no better information than the private sector. debt must be safe. The social planner s problem: max W = U S + U II + U OI = E [µ λ] Rd ι λ (λ) ι α (α) E [υ(λd) + (A R) (Λd)] }{{} real losses from fire sales subject to the same collateral constraint. υ(λd) : fire-sold assets lose value (Shleifer and Vishny, 1992) (A R) (Λd) : some positive NPV projects not undertaken

88 Constrained Effi ciency Constrained effi ciency: no better information than the private sector. debt must be safe. The social planner s problem: max W = U S + U II + U OI = E [µ λ] Rd ι λ (λ) ι α (α) E [υ(λd) + (A R) (Λd)] }{{} real losses from fire sales subject to the same collateral constraint. υ(λd) : fire-sold assets lose value (Shleifer and Vishny, 1992) (A R) (Λd) : some positive NPV projects not undertaken It does not incorporate any q : fire sales are transfer between the agents.

89 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV;

90 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q

91 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q (ii) (Imperfect transparency) SMV of quality is strictly greater than the PMV.

92 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q (ii) (Imperfect transparency) SMV of quality is strictly greater than the PMV.

93 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q (ii) (Imperfect transparency) SMV of quality is strictly greater than the PMV. due to imperfect transparency, only partly internalized through q. q (1 λ) 2α 1 < 1

94 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q (ii) (Imperfect transparency) SMV of quality is strictly greater than the PMV. due to imperfect transparency, only partly internalized through q. q (1 λ) 2α 1 < 1 SMV-PMV is increasing in d.

95 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q (ii) (Imperfect transparency) SMV of quality is strictly greater than the PMV. due to imperfect transparency, only partly internalized through q. q (1 λ) 2α 1 < 1 SMV-PMV is increasing in d. Greater leverage d decreases securitizers stakes in the downturn, causing socially excessive risk-taking.

96 Sources of Ineffi ciency: Imperfect Transparency and Social Marginal Value (SMV) of Quality Proposition (i) (Perfect transparency) SMV of quality λ equals the PMV; Intuition: fully internalized through q (ii) (Imperfect transparency) SMV of quality is strictly greater than the PMV. due to imperfect transparency, only partly internalized through q. q (1 λ) 2α 1 < 1 SMV-PMV is increasing in d. Greater leverage d decreases securitizers stakes in the downturn, causing socially excessive risk-taking. Observation: Transparency and leverage are important indicators on solvency risk.

97 Figure: Difference between the Social and Private Marginal Value of Quality

98 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales

99 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Benefit on funding risk is not internalized: Λ α < 0.

100 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading.

101 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q

102 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q Higher q decreases fire-sale losses.

103 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q Higher q decreases fire-sale losses. q α (1 2λ) f. It is smaller if

104 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q q Higher q decreases fire-sale losses. α (1 2λ) f. It is smaller if cost of transparency is high (e.g., CDO)

105 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q q Higher q decreases fire-sale losses. α (1 2λ) f. It is smaller if cost of transparency is high (e.g., CDO) solvency risk λ is high (e.g., subprime mortgage markets)

106 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q q Higher q decreases fire-sale losses. α (1 2λ) f. It is smaller if cost of transparency is high (e.g., CDO) solvency risk λ is high (e.g., subprime mortgage markets) the price differential f is small (unsophisticated buyers)

107 Social Marginal Value of Transparency Real costs from opacity Λ d }{{} borrowing υ + (A R) }{{} marginal losses from fire sales Λ Benefit on funding risk is not internalized: α < 0. Benefit of transparency is also diversified via trading. But it may be internalized to some extent as transparency may benefit via q q Higher q decreases fire-sale losses. α (1 2λ) f. It is smaller if cost of transparency is high (e.g., CDO) solvency risk λ is high (e.g., subprime mortgage markets) the price differential f is small (unsophisticated buyers) The second and third are particularly important when borrowing d is high.

108 Figure: Private and Socially Optimal Allocation Private Social

109 Implementation : Capital Surcharges on Opaque Securities Instrument : Surcharge on opaqueness τ a and solvency risk τ λ.

110 Implementation : Capital Surcharges on Opaque Securities Instrument : Surcharge on opaqueness τ a and solvency risk τ λ.

111 Implementation : Capital Surcharges on Opaque Securities Instrument : Surcharge on opaqueness τ a and solvency risk τ λ. The socially optimal funding risk does not lead to the socially optimal solvency risk in competitive equilibrium.

112 Implementation : Capital Surcharges on Opaque Securities Instrument : Surcharge on opaqueness τ a and solvency risk τ λ. The socially optimal funding risk does not lead to the socially optimal solvency risk in competitive equilibrium.

113 Implementation : Capital Surcharges on Opaque Securities Instrument : Surcharge on opaqueness τ a and solvency risk τ λ. The socially optimal funding risk does not lead to the socially optimal solvency risk in competitive equilibrium. Proposition Given a constrained-effi cient allocation, there are capital surcharges τ λ and τ α such that the corresponding competitive equilibrium is constrained effi cient.

114 Financial Autarky and Securitization Summary of Securitization (Trading and diversification) Idiosyncratic state-dependent vs independent funding risk (Λ i versus Λ)

115 Financial Autarky and Securitization Summary of Securitization (Trading and diversification) Idiosyncratic state-dependent vs independent funding risk (Λ i versus Λ) Funding risk depends on the transparency of assets produced by others.

116 Financial Autarky and Securitization Summary of Securitization (Trading and diversification) Idiosyncratic state-dependent vs independent funding risk (Λ i versus Λ) Funding risk depends on the transparency of assets produced by others. (Asset pooling) Idiosyncratic state-dependent vs independent asset price (q vs q)

117 Financial Autarky and Securitization Summary of Securitization (Trading and diversification) Idiosyncratic state-dependent vs independent funding risk (Λ i versus Λ) Funding risk depends on the transparency of assets produced by others. (Asset pooling) Idiosyncratic state-dependent vs independent asset price (q vs q) (Asset pooling) Increased pledgeability via diversification ( q vs q)

118 Securitization and Funding Risk Global funding crisis in S = L. Securitization: Λ L = 1. Financial Autarky: E [Λ i ] = 1 [1 λ] }{{} P(α). (2) fraction of good islands

119 Securitization and Funding Risk Global funding crisis in S = L. Securitization: Λ L = 1. Financial Autarky: E [Λ i ] = 1 [1 λ] }{{} P(α). (2) fraction of good islands Securitization spreads out regional risk to other places via diversification

120 Securitization and Funding Risk Global funding crisis in S = L. Securitization: Λ L = 1. Financial Autarky: E [Λ i ] = 1 [1 λ] }{{} P(α). (2) fraction of good islands Securitization spreads out regional risk to other places via diversification Decreases the number of assets yielding more than 1 in the worst state.

121 Securitization and Funding Risk Global funding crisis in S = L. Securitization: Λ L = 1. Financial Autarky: E [Λ i ] = 1 [1 λ] }{{} P(α). (2) fraction of good islands Securitization spreads out regional risk to other places via diversification Decreases the number of assets yielding more than 1 in the worst state. Robustness: Risk-taking investors: E [U(C3 S ) φ] > U (dn).

122 Securitization and Funding Risk Global funding crisis in S = L. Securitization: Λ L = 1. Financial Autarky: E [Λ i ] = 1 [1 λ] }{{} P(α). (2) fraction of good islands Securitization spreads out regional risk to other places via diversification Decreases the number of assets yielding more than 1 in the worst state. Robustness: Risk-taking investors: E [U(C3 S ) φ] > U (dn). It holds as far as the fraction of risk-taking investors is not too large.

123 Figure: Funding crisis in the middle state

124 Securitization and Financial Stability Q j = N.P.V j leverage j funding risk j quality j fire-sale price j. (return to projects ) => leverage => generates two counteracting effects on the MB of quality and transparency:

125 Securitization and Financial Stability Q j = N.P.V j leverage j funding risk j quality j fire-sale price j. (return to projects ) => leverage => generates two counteracting effects on the MB of quality and transparency: 1. MB of quality - fire-sale losses => high quality becomes less valuable.

126 Securitization and Financial Stability Q j = N.P.V j leverage j funding risk j quality j fire-sale price j. (return to projects ) => leverage => generates two counteracting effects on the MB of quality and transparency: 1. MB of quality - fire-sale losses => high quality becomes less valuable. 2. MB of transparency : it decreases fire-sale losses. (1) Funding risk : individual versus aggregate transparency With securitization, this benefit is not internalized.

127 Securitization and Financial Stability Q j = N.P.V j leverage j funding risk j quality j fire-sale price j. (return to projects ) => leverage => generates two counteracting effects on the MB of quality and transparency: 1. MB of quality - fire-sale losses => high quality becomes less valuable. 2. MB of transparency : it decreases fire-sale losses. (1) Funding risk : individual versus aggregate transparency With securitization, this benefit is not internalized. (2) Fire-sale price : With large solvency risk, securitizers make them opaque, q α (1 2λ) < 0.

128 Figure: Securitization and Financial Autarky 0.5 Securitization Autarky

129 Summary Benefit of transparency is diversified. Securitizers do not fully internalize the benefit of it.

130 Summary Benefit of transparency is diversified. Securitizers do not fully internalize the benefit of it. A decrease in transparency further exacerbates solvency risk.

131 Summary Benefit of transparency is diversified. Securitizers do not fully internalize the benefit of it. A decrease in transparency further exacerbates solvency risk. Solvency and funding risks are high particularly when borrowing is high.

132 Key Objective : Profits = N.P.V quality leverage funding risk fire-sale price. }{{} securitizers losses from funding crises Social Welfare = N.P.V E [υ(λd) + (A R) (Λd)] }{{} real losses from fire sales