Macroeconomics and Financial Markets. Housing and Credit Markets: Bubbles and Crashes

Transcription

1 Macroeconomics and Financial Markets... or, better,... Housing and Credit Markets: Bubbles and Crashes Veronica Guerrieri 1 and Harald Uhlig 2 1 University of Chicago Booth School of Business veronica.guerrieri@chicagobooth.edu 2 University of Chicago Department of Economics huhlig@uchicago.edu February 26, 2016 Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

2 Outline 1 Introduction 2 Catastrophes 3 Conclusions Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

3 Introduction Outline 1 Introduction 2 Catastrophes 3 Conclusions Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

4 Introduction Goal The purpose of this chapter is to examine a key connection between financial markets and macroeconomic activity. Specifically, we wish to examine a connection between two features 1 A boom-bust in house prices. 2 A boom-bust in credit markets. Large and growing literature, exploring these separately. This chapter: connection. Two possible channels: 1 Credit boom-bust => House price boom-bust. 2 House price boom-bust => credit boom-bust. Aggregate repercussions: not this chapter. Issues are far from resolved. Encourage future research. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

5 Introduction The S&P/Case-Shiller Home Price Indices Source: Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

6 Introduction Subprime Mortgage Originations Source: The Financial Crisis Inquiry Report, National Commission, January 2011 Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

7 Outline Introduction 1 Several small models to illustrate key ideas. 2 To set the stage: a stark model. Banks, HH with mortgages, interaction. Given: dynamics of both leverage and house prices. Keep unconstrained house prices constant. Study the response to a boom and bust in leverage and its implications for credit-constrained house-prices. Keep leverage constant. Study response to a boom and bust in house prices to banks. 3 Catastrophe Models: credit boom-bust. Savings glut leads to crash in lending and crash in house prices. 4 Bubbles: house price boom-bust. 1 Bubbles, type 1. Dynamically inefficient economies, bubbles do not grow or even die out on their own. 2 Bubbles, type 2. Dynamically efficient. Buyers hope to find greater fools. 5 (Some) evidence Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

8 Catastrophes Outline 1 Introduction 2 Catastrophes 3 Conclusions Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

9 Catastrophes Saving Glut focus on the boom and bust in the mortgage market idea: increase in credit availability ( saving glut ) can endogenously generate a credit cycle in spirit of Boissay, Collard, and Smets (2016, JPE, forthcom.) mechanism: as more credit is available, worse borrowers get funding and at some point good borrowers step out from the market generating a credit crush key: adverse selection can generate multiple equilibria growing body of literature focus on leveraging/deleveraging cycle of households sector... Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

10 Catastrophes Model two periods t = 1, 2 continuum of 2 types of agents: households and banks households utility in period 2 u(c, h) = c +γh. houses in fixed size h h {0, h} house price fixed to 1 households get income draw y in period 2 and decide to buy a house in period 1 if they buy a house they need to borrow 1 if borrow 1 at price p, have to repay 1/p in period 2 if hh defaults: she looses house, pays cost δ, lender gets zero. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

11 Catastrophes Equilibrium cutoffs Two-dimensional idiosynchratic risk: [ν, y]. ν [0, 1] distributed according to G(ν): household type Given ν, households income y distributed acc. to F ν (y) F ν2 (y) FOSD F ν1 (y) when ν 1 < ν 2 ν is household s private information at the beginning of period 1 households choose 1 whether to pay k to verify their type: v(ν) {0, 1} 2 whether to buy a house/borrow: h(ν) {0, h} h(ν) = v(ν) = 1 ν can borrow at price p(ν) h(ν) = 1 and v(ν) = 0 ν can borrow at pooling price p P Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

12 Catastrophes Default household with realized income y would like to repay if y 1/p+γ h y δ we choose δ large enough so that households would like to pay back if they can let χ(y, p) be the default indicator: χ(y, p) = 1 iff y 1/p, otherwise default define π(ν, p) E[χ(y, p) = 1 ν, p] = 1 F ν ( 1 p ) Proposition π(ν, p) is increasing in ν and increasing in p better types have higher repayment probability for given p. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

13 Lending Market Catastrophes banks can borrow at some exogenous rate R no arbitrage gives us: ( ) 1 F 1 ν p(ν) p(ν) = R [ ( ) ] E 1 F 1 ν ν [ν(p P ), ν(p P )] p P p = P R where [ν(p P ), ν(p P )] is the set of hh borrowing at p P Proposition If it exists, the type contingent contract p(ν) is increasing in ν and decreasing in R. What about p P? More complicated. We do not have a tight characterization there. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

14 Households problem Catastrophes Utility of household ν can be written as { } Ū(ν, p P ) = max U B (ν, p P ), U V (ν), U N (ν), where U B (ν, p P ) = 1 p P (y U V (ν) = U B (ν, p(ν)) k U N (ν) = ydf ν (y) 1 1 p P +γ h)df p ν (y)+ P (y δ)df ν (y) 0 Proposition Suppose dfν(x)/dν 1 F ν(x) is increasing in x for all ν. Then, there exists a value p so that for all p P p there are two cutoffs ν(p P ) and ν(p P ) such that h(ν) = 0 iff ν < ν(p P ) and v(ν) = 1 iff ν ν(p P ). Moreover, ν(p P ) and ν(p P ) are respectively decreasing and increasing in p P. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

15 Households problem Catastrophes U V - U N U B P )- U N 0 (p P ) (p P ) Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

16 Catastrophes Multiple Equilibria there may be two types of equilibria: 1 pooling equilibrium: nobody verifies his type pooling price high no incentive to verify 2 separating equilibrium: good types verify pooling price low incentive to verify if good enough saving glut can generate a (comp.stat.) credit crash: when R large, costly to borrow bad households do not borrow at all nobody verifies as R declines pooling price increase more bad households borrow dampening the increase in price as R keeps declining good types start verifying pooling price decrease discretely = catastrophe as R increases back from there, we may remain stuck on the low-pooling-price, low-lending activity equilibrium branch. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

17 Catastrophes The Equilibrium Manifold 0.64 Equilibrium Manifold p R Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

18 Numerical examples Catastrophes Tricky! Two cases. For both:y {0; ȳ}. ν = P(y = ȳ). γ h = 2,δ = 0.1. Case 1: mix expon. density and normal density, trunc.: h(ν) ω λe λν +(1 ω)e (ν νe)2/(2σ2) 1 e λ 2πσ Params: κ = 0.25,λ = 20,ν e = 0.1,σ = 0.2,ω = 0.6 Equil: R = 1.3 uniq sep,r = 1.4 mult,r = 1.5 uniq pool. Case 2: A mixture of two exponential densities, h(ν) = ω λ 1e λ 1ν 1 e λ 1 +(1 ω) λ 2e λ2ν 1 e λ 2 Params: κ = 0.15,λ 1 = 20,λ 2 = 5,ω = 0.8. Equ: R = 1.4 uniq sep,r = 1.58 mult,r = 1.65 uniq pool. Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

19 Case 1: type density Catastrophes 14 Density for ν Density ν Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

20 Catastrophes Case 1: particip.choices, bounds ν(p P ) and ν(p P ) 1 Bounds for R=1.30,(blue:low,red:high) ν p Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

21 Catastrophes Case 1: finding the fixed points 0.7 Fixed point calculation E[π]/R: Two Point Example E[π]/R p Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

22 Catastrophes Case 1: Lending Activity Volume lent Volume lent at R and p: Two Point Example p Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

23 Catastrophes Case 1: The Equilibrium Manifold 0.64 Equilibrium Manifold p R Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

24 Case 2: type density Catastrophes 16 Density for ν Density ν Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

25 Catastrophes Case 2: particip. choices, bounds ν(p P ) and ν(p P ) 1 Bounds for R=1.40,(blue:low,red:high) ν p Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

26 Catastrophes Case 2: finding the fixed points 0.62 Fixed point calculation E[π]/R: Two Point Example 0.6 E[π]/R p Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

27 Catastrophes Case 2: Lending Activity Volume lent at R and p: Two Point Example Volume lent p Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

28 Catastrophes Case 2: The Equilibrium Manifold 0.9 Equilibrium Manifold p R Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

29 Conclusions Outline 1 Introduction 2 Catastrophes 3 Conclusions Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30

30 Conclusions Summing Up key interplay between the housing market cycle and the credit market cycle large literature focused on one of the two possible stories: 1 credit market boom-bust housing market boom-bust 2 housing market boom-bust credit market boom-bust we first described a stark model to think about this interplay then we developed two models inspired by existing literature: 1 given prices, increase in credit availability generates a credit boom-bust 2 given leverage, irrational optimism can generate a house price boom-bust more work required to connect the two... Guerrieri-Uhlig (University of Chicago) Housing and Credit: Bubbles and Crashes February 26, / 30