From the ADEMU project series of lectures, Greg Kaplan, Kurt Mitman and Gianluca Violante examine the property boom-bust, and ask whether it could have been cushioned by a debt-forgiveness policy. Taken from the New Developments in Macroeconomics lecture at UCL London, November 2016

1. Consumption and House Prices
in the Great Recession: Model Meets Evidence
Greg Kaplan
Kurt Mitman
Gianluca Violante
UCL
November 9, 2016

2. Three questions
1. What shock(s) drove the boom-bust in ph?
• Credit conditions vs expectations about future growth in ph
1

3. Three questions
1. What shock(s) drove the boom-bust in ph?
• Credit conditions vs expectations about future growth in ph
2. Why the corresponding boom-bust in C?
• Channels: Collateral vs wealth effects
1

4. Three questions
1. What shock(s) drove the boom-bust in ph?
• Credit conditions vs expectations about future growth in ph
2. Why the corresponding boom-bust in C?
• Channels: Collateral vs wealth effects
3. Could a debt-forgiveness policy have cushioned the bust?
• Study large-scale Principal Reduction program
1

5. Methodology
• Model: aggregate shocks move equilibrium ph
2

6. Methodology
• Model: aggregate shocks move equilibrium ph
• Parameterize: match cross-sectional and lifecycle micro data
2

7. Methodology
• Model: aggregate shocks move equilibrium ph
• Parameterize: match cross-sectional and lifecycle micro data
• Simulate boom-bust
• Compare against aggregate time-series data
• House prices
• Consumption
• Rent-price ratio
• Home ownership
• Leverage
• Foreclosures
• Compare against micro data
2

8. Methodology
• Model: aggregate shocks move equilibrium ph
• Parameterize: match cross-sectional and lifecycle micro data
• Simulate boom-bust
• Compare against aggregate time-series data
• House prices
• Consumption
• Rent-price ratio
• Home ownership
• Leverage
• Foreclosures
• Compare against micro data
• Counterfactuals to address our questions
2

9. Preview of main results
1. Sources of boom-bust in ph and C
• Main driver is beliefs, not change in credit conditions
• Credit conditions important for ownership, leverage and
foreclosure
3

10. Preview of main results
1. Sources of boom-bust in ph and C
• Main driver is beliefs, not change in credit conditions
• Credit conditions important for ownership, leverage and
foreclosure
2. Transmission mechanism to household consumption
• Mostly a wealth effect, not collateral effect
3

11. Preview of main results
1. Sources of boom-bust in ph and C
• Main driver is beliefs, not change in credit conditions
• Credit conditions important for ownership, leverage and
foreclosure
2. Transmission mechanism to household consumption
• Mostly a wealth effect, not collateral effect
3. Effectiveness of mortgage modification program
• Big effect on foreclosures, but negligible impact on ph and C
3

12. Model

13. Model
Demographics
• OLG lifecycle economy with work & retirement
Endowments
• Workers face uninsurable risk in individual earnings y
Preferences
• Utility over nondurable c and housing services h
Housing
• Finite number of house sizes h ∈ H
• Households can buy a unit of h at price ph, or rent it at rate ρ
• Linear transaction cost κh · (phh) for sellers
4

14. Financial instruments
Liquid saving (b > 0): one-period bond, exogenous interest rate rb (fixed)
5

15. Financial instruments
Liquid saving (b > 0): one-period bond, exogenous interest rate rb (fixed)
Mortgages (m): long-term, fixed-rate debt contract
• Price schedule qj (h, m, b, y) set by competitive banking sector
• Amortized over remaining lifetime at rate rb (1 + ι)
• Refinancing option available (cash-out) at cost κm
• Max Loan-to-Value at origination only m ≤ λm
phh
• Max Payment-to-Income at origination only π ≤ λπ
y
5

16. Financial instruments
Liquid saving (b > 0): one-period bond, exogenous interest rate rb (fixed)
Mortgages (m): long-term, fixed-rate debt contract
• Price schedule qj (h, m, b, y) set by competitive banking sector
• Amortized over remaining lifetime at rate rb (1 + ι)
• Refinancing option available (cash-out) at cost κm
• Max Loan-to-Value at origination only m ≤ λm
phh
• Max Payment-to-Income at origination only π ≤ λπ
y
Foreclosure
• Default on mortgage debt: incur a utility loss
5

17. Financial instruments
Liquid saving (b > 0): one-period bond, exogenous interest rate rb (fixed)
Mortgages (m): long-term, fixed-rate debt contract
• Price schedule qj (h, m, b, y) set by competitive banking sector
• Amortized over remaining lifetime at rate rb (1 + ι)
• Refinancing option available (cash-out) at cost κm
• Max Loan-to-Value at origination only m ≤ λm
phh
• Max Payment-to-Income at origination only π ≤ λπ
y
Foreclosure
• Default on mortgage debt: incur a utility loss
HELOCs (b < 0)
• One-period borrowing (b ≥ −λbphh), at rate rb (1 + ι), non-defaultable
5

18. Closing the model
Final good sector
• Y = Z ¯N → w = Z
Construction sector
• Labor + housing permits → aggregate housing investments
I(ph)
Rental sector
• Buys housing from sellers and rents them out, or vice-versa,
sells rental units to home buyers
• Operating cost ψ per unit of housing owned and rented out
• Zero-profit condition yields equilibrium rental rate ρ
Government
• Taxes workers (with mortgage interest deduction) and
properties, sells land permits, and pays SS benefits to retirees
6

19. Aggregate shocks
1. Aggregate labor income: Z
2. Credit conditions: (i) credit limits (λm, λb, λπ)
(ii) intermediation wedge ι
7

20. Aggregate shocks
1. Aggregate labor income: Z
2. Credit conditions: (i) credit limits (λm, λb, λπ)
(ii) intermediation wedge ι
3. Beliefs / News about future housing demand:
Three regimes for ϕ (share of housing services in u):
(a) ϕL: low housing share and unlikely transition to ϕH
(b) ϕ∗
L: low housing share and likely transition to ϕH
(c) ϕH: high housing share
Boom-Bust: shift from (a) to (b), and back to (a)
7

21. Parameterization

22. Parameterization strategy
Parameter values disciplined by facts from household-level
micro-data
Age
25 30 35 40 45 50 55 60 65 70 75 80
HomeOwnershipRate
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Model
Data
Age
30 40 50 60 70 80
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Leverage - Model
Leverage - Data
Frac Borrowing - Model
Frac Borrowing - Data
• Life-cycle profiles of ownership and leverage
8

23. Parameterization strategy
Parameter values disciplined by facts from household-level
micro-data
• Distributional stats: mortgages, housing wealth, renters, and
consumption
Moment Empirical value Model Value
Fraction homeowners w/ mortgage 0.66 0.57
Aggr. mortgage debt / housing value 0.42 0.36
P10 LTV ratio for mortgagors 0.15 0.14
P50 LTV ratio for mortgagors 0.57 0.59
P90 LTV ratio for mortgagors 0.92 0.92
Aggr. home-ownership rate 0.66 0.67
P10 house value / earnings 0.90 1.0
P50 house value / earnings 2.1 2.0
P90 house value / earnings 5.5 4.5
Avg.-size owned house / rented 1.5 1.4
Avg. earnings owners / renters 2.05 2.02
BPP consumption insurance coef 0.36 0.43
8

24. Shock Processes
1. Aggr. labor income: NIPA wages & salaries per capita
2. Credit conditions: λm: 95% → 110%, λb: 20% →30%
λπ : 25% → 60%, ιm : 100 BP → 75 BP
3. Beliefs: Case-Shiller-Thompson &
Burnside-Eichenbaum-Rebelo
9

25. Shock Processes
1. Aggr. labor income: NIPA wages & salaries per capita
2. Credit conditions: λm: 95% → 110%, λb: 20% →30%
λπ : 25% → 60%, ιm : 100 BP → 75 BP
3. Beliefs: Case-Shiller-Thompson &
Burnside-Eichenbaum-Rebelo
Realized path for shocks
Year
2000 2005 2010 2015
0.96
0.98
1
1.02
1.04
Productivity, Z
Year
2000 2005 2010 2015
0.9
0.95
1.0
1.05
1.1
Financial Deregulation, 6
m
Year
2000 2005 2010 2015
0.02
0.03
0.04
0.05
0.06
0.07
Expected Price Growth
• The shift in beliefs hits in 2001 and reverts back in 2007
9

26. Q1
What caused the boom-bust in ph and C?

27. Consumption and house price dynamics
Year
2000 2005 2010 2015
0.8
0.9
1
1.1
1.2
1.3
House Price
Model
Data
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Consumption
10

28. Consumption and house price dynamics
Year
2000 2005 2010 2015
0.8
0.9
1
1.1
1.2
1.3
House Price
Benchmark
Belief Only
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Consumption
10

29. Consumption and house price dynamics
Year
2000 2005 2010 2015
0.8
0.9
1
1.1
1.2
1.3
House Price
Benchmark
Belief Only
Income Only
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Consumption
10

30. Consumption and house price dynamics
Year
2000 2005 2010 2015
0.8
0.9
1
1.1
1.2
1.3
House Price
Benchmark
Belief Only
Income Only
Credit Only
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Consumption
10

31. Beliefs vs actual change in preferences
11

32. Beliefs vs actual change in preferences
Year
2000 2005 2010 2015
0.8
0.9
1
1.1
1.2
1.3
House Price
Benchmark
Demand Only
Year
2000 2005 2010 2015
0.9
0.95
1
1.05
1.1
Consumption
• Preference shock: similar rise in ph, but C falls!
11

33. Dynamics of rent-price ratio
Year
2000 2005 2010 2015
0.7
0.8
0.9
1
1.1
Benchmark
12

34. Dynamics of rent-price ratio
Year
2000 2005 2010 2015
0.7
0.8
0.9
1
1.1
Benchmark
Belief Only
Income Only
Credit Only
ρ = ψ + ph −
(
1 − δh − τh
1 + rb
)
Eph
[
p′
h
]
• Belief about future appreciation essential
12

35. Dynamics of home ownership
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Benchmark
13

36. Dynamics of home ownership
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Benchmark
Belief Only
Income Only
Credit Only
• Loosening of credit limits drives rise in home-ownership
13

37. Change in home ownership by age: data and model
Age
30 40 50 60
Log-change(relativetomean)
-0.05
0
0.05
0.1
0.15
Boom
Data
Model
Age
30 40 50 60
Log-change(relativetomean)
-0.15
-0.1
-0.05
0
0.05
Bust
Data
Model
• It’s the young who go in/out of housing market
14

38. Dynamics of leverage and foreclosure
Year
2000 2005 2010 2015
0.8
1
1.2
1.4
1.6
1.8
Leverage
Year
2000 2005 2010 2015
0
0.01
0.02
0.03
0.04
Foreclosure rate
Benchmark
15

39. Dynamics of leverage and foreclosure
Year
2000 2005 2010 2015
0.8
1
1.2
1.4
1.6
1.8
Leverage
Year
2000 2005 2010 2015
0
0.01
0.02
0.03
0.04
Foreclosure rate
Benchmark
Belief Only
Income Only
Credit Only
• Credit loosening is key for constant leverage pre-boom
• Interaction belief-credit important for foreclosure
15

40. Why credit shock does not affect ph
• Max LTV/PTI ratios affect housing demand if renters (extensive
margin) or home-owners (intensive margin) are constrained in
housing choice
1. UP: Rental market relaxes these constraints
2. DOWN: Long-term mortgage debt relaxes these constraints
16

41. Why credit shock does not affect ph
• Max LTV/PTI ratios affect housing demand if renters (extensive
margin) or home-owners (intensive margin) are constrained in
housing choice
1. UP: Rental market relaxes these constraints
2. DOWN: Long-term mortgage debt relaxes these constraints
• Are we missing the ‘credit supply’ aspect of the shock, i.e.
cheap credit flowing to low-quality borrowers?
1. Endogenous relaxation in lending standards in response to
belief-driven boom
16

42. Cheaper credit for ‘low-quality’ borrowers
Leverage
0.7 0.8 0.9 1 1.1
EndogenousBorrowingRate(1/q
m
-1)
0.04
0.06
0.08
0.1
0.12
0.14
No shocks
Credit Only
All Shocks
• Also lenders expect prices to rise and default rates to fall
17

43. Model Meets (New) Micro Evidence

44. New narrative I
• Mian-Sufi: credit growth concentrated in low-income groups
• Foote et al.: no, equally distributed across income groups
18

45. New narrative I
• Mian-Sufi: credit growth concentrated in low-income groups
• Foote et al.: no, equally distributed across income groups
0.1.2.3.4.5
ShareofDebt
1 2 3 4 5
Income Quintile of Household
Shares of Mortgage Debt
2001 2007
• Low-income hh switch from rent to buy, high-income hh upsize
18

46. New narrative II
• Mian-Sufi: mortgage origin. concentrated in subprime groups
• Adelino et al.: no, equally distributed across groups
19

47. New narrative II
• Mian-Sufi: mortgage origin. concentrated in subprime groups
• Adelino et al.: no, equally distributed across groups
0.2.4.6.8
ShareofDebt
Above Median Below Median
Default Risk
Shares of Originated Mortgage Debt
2001 2007
• Young hh switch from rent to buy, older hh upsize
19

48. New Narrative III
• Mian-Sufi: foreclosures concentrated in subprime groups
• Albanesi et al.: no, proportionally rising more for other groups
20

49. New Narrative III
• Mian-Sufi: foreclosures concentrated in subprime groups
• Albanesi et al.: no, proportionally rising more for other groups
0.2.4.6.81
ShareofForeclosures
2004 2006 2008 2010 2012
Year
IncomeQuintile=1 IncomeQuintile=2
IncomeQuintile=3 IncomeQuintile=4
IncomeQuintile=5
• Everyone levers up, including middle-income households
20

50. Q2
How does the fall in ph transmit to C?

51. Deleveraging or wealth effect in the bust?-.3-.2-.10.1
ChangeinLogConsumption
0 .05 .1 .15
Debt as a Fraction of Total Wealth
Renters Owners
-.3-.2-.10.1
ChangeinLogConsumption
0 .1 .2 .3 .4
Housing Share of Total Wealth
Renters Owners
Deleveraging: WEAK Wealth effect: STRONG
21

52. Deleveraging or wealth effect in the bust?-.3-.2-.10.1
ChangeinLogConsumption
0 .05 .1 .15
Debt as a Fraction of Total Wealth
Renters Owners
-.3-.2-.10.1
ChangeinLogConsumption
0 .1 .2 .3 .4
Housing Share of Total Wealth
Renters Owners
Deleveraging: WEAK Wealth effect: STRONG
• Consistent with Kaplan-Mitman-Violante (2016): ’Non-durable
Consumption and Housing Net Worth in the Great Recession:
Evidence form Easily Accessible Data’
21

53. Q3
Could a massive debt forgiveness program
have cushioned the bust?

54. Counterfactual principal reduction program
All homeowners with LTV >95%: forgive excess debt
22

55. Counterfactual principal reduction program
All homeowners with LTV >95%: forgive excess debt
Year
2000 2005 2010 2015
0.8
1
1.2
House Price
Bench.
Mod.
Year
2000 2005 2010 2015
0.95
1
1.05
1.1
Consumption
Year
2000 2005 2010 2015
0
0.01
0.02
0.03
Foreclosure rate
Year
2000 2005 2010 2015
0.8
1
1.2
1.4
1.6
Leverage
• Beneficiaries account for small share of C + do not foreclose
22

56. Summary: what did we learn from the model?
1. Shift in expected house appreciation key to boom-bust in ph
2. This explanation is consistent with recent micro evidence
3. ∆ph transmits to ∆C through wealth effects
4. Principal reduction program would not have mitigated drop in C
23

57. Summary: what did we learn from the model?
1. Shift in expected house appreciation key to boom-bust in ph
2. This explanation is consistent with recent micro evidence
3. ∆ph transmits to ∆C through wealth effects
4. Principal reduction program would not have mitigated drop in C
Thanks!
23