Post on 10-Jan-2016
description
transcript
Household Debt and Credit Constraints: Comparative Micro Evidence From Four
OECD Countries
Jonathan Crook (U Edinburgh)
Stefan Hochguertel (VU Amsterdam)
od1
Introduction
• Debt holding has increased over last 15 yrs
• Empirical literature on liquidity constraints has not come up yet with internationally comparative figures: how important are they?
• Arguably, importance of both debt holding and constraints differs across countries
• Institutional differences between countries will matter
od2
Household Debt in OECD Countries============================================== Debt/YD Mortg/GDP 1995 2005 % 1992 2002 %----------------------------------------------DK 112.9 155.2 37.5 63.9 74.3 16.3NL 63.4 134.1 111.5 40.0 78.8 97.0PORT 46.8 112.6 140.6 12.8 49.8 289.1US 78.8 111.1 41.0 45.3 58.0 28.0SP 47.4 93.5 97.3 11.9 32.3 171.4GER 74.3 83.2 12.0 38.7 54.0 39.5SW 54.7 78.3 43.1 37.5 40.4 7.7FR 47.8 65.2 36.4 21.0 22.8 8.6BEL 45.7 54.2 18.6 19.9 27.9 40.2FIN 47.2 58.6 24.1 37.2 31.8 -14.5GR 8.6 44.9 422.1 4.0 13.9 247.5IT 24.6 43.1 75.2 6.3 11.4 81.0==============================================Source: OECD
od3
This Paper
• Use micro data from four OECD countries (Italy, US, NL, Spain) and estimate equations for ‘demand for debt’ and ‘credit constraints’
• Consider various selection issues
• Provide comparable estimates and relate this, where possible, to differential incentives and constraints arising out of differing institutional designs between countries
od4
• Find: pronounced differences as regards effects of incomes, net worth, age, household composition
• Selection effects of various kinds appear unimportant for debt holding equations
• DYYP [= difference between current and ‘permanent income’] has differential effects for credit applications, credit constraints, and conditional debt holding:
– applications: “–” NL, “0” IT, US, SP
– constraints: “0” NL, “+” IT, US, “0” SP
– debt holding: “0” IT, “–” NL, US, “0” SP
od5
Simulation Results
Assume
• CRRA utility function; R=1; bequests=0.
• Individuals retire at fixed date tR when income drops to fraction of last earned income
• Permanent income follows an AR(1) process, grows at rate G, subject to permanent shocks
• Current income subject to transitory shocks
• Solution method follows Deaton 1991
od13
_____________________________________________ Benchmark (Table 1) age 30 40 50 60 70 ______________________________________________ holds any debt % 81.1 73.4 23.3 0.2 0.0 constrained % 17.6 2.5 0.0 0.0 0.0 avg. debt/avg. income 0.197 0.235 0.041 0.000 0.000 indiv. debt/income mean 0.207 0.267 0.058 0.001 0.000 (unconditional) median 0.183 0.200 0.000 0.000 0.000 stddv 0.183 0.281 0.152 0.014 0.000 debt/income mean 0.256 0.364 0.247 0.312 0.000 (if debt > 0) median 0.233 0.312 0.185 0.312 0.000 stddv 0.170 0.269 0.230 0.018 0.000 _____________________________________________________________________
Generalisations
Debt/Y hump shaped and reaches peak at age 40
Debt incidence is monotonically decreasing with age
If cash on hand< 110% of consumption at given age credit constrained then:
• 18% constrained at age 30, 2.5% at age 40, 0% thereafter
• If decrease replacement rate fraction constrained decreases at all ages
• If increase growth rate of permanent income fraction constrained decreases at all ages
• If increase time preference rate fraction constrained increases at all ages
• If increase interest rate fraction constrained decreases at all ages
od13b
Empirical Literature on Debt and Constraints
International comparisons
• Jappelli & Pagano 1989 AER Bacchetta & Gerlach 1997 JME
Micro Data
US
• Jappelli 1990 QJE Cox & Jappelli 1993 JMCB
• Duca & Rosenthal 1993 JFinInterm Crook 1996 & 2001 App Fin Ec
• Gropp, Schultz & White 1997 QJE Jappelli et al 1998 REStats
• Ferri & Simon 2002 UB WP Lyons 2003 JCA
• Grant 2005 EUI WP
Italy Australia
• Fabri & Padula 2004 JBF La Cava & Simon BoA WP 2003
• Magri 2002 BoI WP
od14
• Income insurance (SS, pensions, UI, EPL)
• Bankruptcy
• Usury
• Judicial efficiency, information sharing
• Asset prices (homes), homeownership
• Mortgage market institutions
• Taxation of incomes and wealth
• Preferences
• Organization of financial markets
Institutional Aspects
od14b
Institutional Differences
NL IT US Sp
Unemployment benefit insurance against income shocks greater in NL and Sp than IT and US
Unemployment spend/GDP (%) 2.3 0.9 0.4 2.2
Duration of unemployment benefit 60 6 6 24(months)
Employment protection 2.3 2.4 0.7 3.7(OECD index 0 to 6)(rank out of 28) 12 11 ? 4-----------------------------------------------------------------Bankruptcy protection higher in US than IT and NL
Bankruptcy discharge repaym no yes possible plan
----------------------------------------------------------------
od15
-------------------------------------------------------------------Judicial efficiency and information sharing between lenders highest in NL and US than IT and SP NL IT US Sp
Judicial enforcement
days to collect bounced check 39 645 54 147days to evict delinquent tenant 52 630 49 183
Time to repossess (months) 6 60-84 8 7-9
#reports issued/person 0.64 0.046 2.3 naby private credit bureau.
-----------------------------------------------------------------
od15b
Homeownership (%) 1990 45 68 64 782002 53 80 68 85
House price increases in 90’s rapid decline rapid rapid
Home equity withdrawal high none high none
-----------------------------------------------------------------
NL IT US Sp---------------------------------------------------------------------------------------------------------------------Highest downpayment % and lowest LTV values in IT, mid values for NL and SP, lowest downpayment % for US
Downpayment (%) 25 40 11 20
Loan to value Typical(%) 90 55 78 70Max (%) 115 80 na 100
Typical term (years) 30 15 30 15
----------------------------------------------------------------Tax deductibility higher in NL and US than IT and SP
Tax deductibility for mortgageon main residence Yes capped capped capped
C20% of $100k Є9k interest
---------------------------------------------------------------Italian banks are higher cost that elsewhere. Competition in all European markets increasing due to entry.
od16
Summary-----------------------------------------------------
Greatest Min CreditIncome insurance NL, Sp > IT, US IT, US
Bankruptcy US > IT, NL IT, NL
Usury na
Judicial efficiency, NL, US> IT, SP IT, SPinformation sharing
Asset prices inflation(homes), NL, US > IT, SPIT, SP
Homeownership IT, SP > NL, USNL, US
Downpayment IT, NL > US, SPIT, NL
Taxation deduction NL, US > IT, SP IT, SP
Organization of financial Markets (Bank costs) IT > US, NL, SPIT-------------------------------------------------------------------
od16b
Micro Data
• US: Survey of Consumer Finances, Federal Reserve Board, SCF
• Italy: Survey of Household Income and Wealth, Bank of Italy, SHIW
• Netherlands: DNB Household Survey, Central Bank of the Netherlands, DHS
• Spain: Spanish Survey of Household Finances, Bank of Spain, EFF
od17
US Data
• SCF: 1992, 1995, 1998, 2001• Repeated cross section; household level• Triennial; 4000-4500 hh per year• Large variety of questions on wealth and debt holding (account-
level)• Oversampling of high wealth households• Multiple imputations
od18
Italian Data
• SHIW: 1991, 1993, 1995, 1998, 2000, 2002, 2004• Cross section with panel component (> 40%); household level• Biennial; 5000-8000 hh per year• Large variety of questions on wealth and debt holding (account-
level)• Some missing values imputed, only one implicate
od19
Dutch Data
• 1993-2004 DHS• Panel data; household level• Annual frequency; 2000-3000 hh per year• Large variety of questions on wealth and debt holding (account-
level)
od20
Spanish Data
•2002 EFF•Cross section household level•5000 hh•Large variety of questions on wealth and debt
holding (account level)•Oversampling of wealthy•Multiple imputations
od21
Permanent income
od22
Incidence of Debt================================================================ Mortgages Other Total NL IT Sp US NL IT Sp US NL IT SP US-----------------------------------------------------------------------1991 11.0 14.1 23.0 1992 41.8 64.7 73.51993 40.6 12.5 46.0 15.0 64.7 25.11994 39.2 43.6 64.11995 41.2 13.4 43.4 43.3 14.1 66.2 64.6 24.6 74.71996 42.9 44.4 66.51997 43.3 43.9 66.31998 43.6 9.1 45.3 43.5 16.4 63.7 66.8 22.9 74.31999 42.1 42.8 67.22000 45.3 9.2 41.3 16.5 68.3 23.12001 42.6 46.6 40.0 64.2 65.6 75.52002 43.6 10.2 26.7 41.3 13.9 24.4 67.0 21.4 43.6 2003 41.0 40.5 65.42004 42.4 11.9 36.2 15.0 65.2 23.5=======================================================================NL: DHS, IT: SHIW, SP: EFF, US: SCF
od23
Mean Debt Holding (1992 Euros, 1000's)================================================================ Mortgages Other Total NL IT Sp US NL IT Sp US NL IT Sp US----------------------------------------------------------------1991 1.75 1.44 3.19 1992 35.37 6.82 42.191993 29.57 2.18 3.75 1.72 33.32 3.90 1994 26.61 3.28 29.90 1995 27.66 2.25 35.51 3.45 1.65 7.65 31.11 3.90 43.161996 29.00 3.43 32.42 1997 27.91 3.26 31.20 1998 25.77 1.66 41.28 2.77 2.92 10.35 28.54 4.58 51.621999 27.28 2.59 29.86 2000 27.45 2.21 3.24 2.38 30.11 4.59 2001 33.63 45.24 3.90 9.55 37.90 54.792002 36.76 2.56 8.26 3.21 2.01 2.04 40.23 4.56 10.30 2003 34.18 4.21 38.02 2004 35.20 3.73 3.99 2.31 39.13 6.04 ================================================================NL: DHS, IT: SHIW, SP: EFF, US: SCF
od24
Median Debt Holding (1992 Euros, 1000's)if positive================================================================ Mortgages Other Total NL IT Sp US NL IT Sp US NL IT Sp US----------------------------------------------------------------1991 10.86 4.89 5.97 1992 53.15 5.40 21.261993 59.90 11.37 2.72 3.21 34.49 6.43 1994 57.36 2.65 30.96 1995 55.99 11.30 56.55 2.48 3.16 6.20 32.30 6.87 24.141996 56.58 2.55 33.99 1997 53.76 2.07 29.57 1998 49.36 12.55 64.05 2.23 4.18 7.78 28.32 6.27 33.961999 51.57 2.38 29.75 2000 43.08 16.04 2.33 4.01 18.05 6.42 2001 61.13 70.18 3.71 7.86 37.12 36.962002 64.44 18.44 24.46 3.08 4.43 4.29 36.56 7.38 16.01 2003 67.90 2.70 35.92 2004 68.01 24.37 2.76 4.21 34.76 8.78 ================================================================NL: DHS, IT: SHIW, SP: EFF, US: SCF
od25
Applicants and Rejections============================================================== % Apply % Reject % Reject|Apply NL IT SP US NL IT SP US NL IT SP US--------------------------------------------------------------1991 0.9 1992 22.5 1993 22.2 0.8 1.1 4.31995 19.9 5.6 63.6 0.9 0.9 20.4 4.4 16.2 32.01998 21.4 6.0 63.6 0.8 0.5 21.8 3.9 7.7 34.22000 21.9 5.4 1.9 0.4 5.6 8.02001 26.1 64.9 1.5 19.9 3.7 30.72002 24.9 4.2 20.8 2.5 0.5 1.1 9.1 11.7 5.1============================================================== ================================= %Rejected or Discouraged NL IT SP US --------------------------------------------------------- 1993 2.3 3.0 1995 2.9 2.3 28.6 1998 3.0 2.8 28.4 2000 3.1 1.7 2001 2.3 26.9 2002 3.5 2.2 3.4 ================================= NL: DHS, IT: SHIW, SP, EFF, US: SCF
od26
Observational regimes
od27
Selection mechanisms
• Debt holding
• Wants debt
• Applies
• Accepted
od28
Empirical modeling
• Observability rule:
• ML (normality, random effects, simulation)• Various versions• Non-convergence, partial convergence, etc
od29
• SHIW: no random effects in selection equations, but also no selection
• DHS: random effects, but no selection effects
• EFF: cross section, Tobit
• SCF: pooled cross section, no selection effects found (consistent two- stage estimator with double selection rule)
Shall focus on single equation models for comparability
od30
Prob(Credit Application) – marginal effects=========================================================== NL IT US SP -----------------------------------------------------------wealth -0.0064** -0.0030** -0.0023* -0.0172** income1 0.0037 0.0033 0.0212** 0.0361income2 0.1316** 0.0516** 0.2897** 0.1050*income3 0.2743** -0.0311 0.0729 0.0908income4 0.0164 0.0211 0.0332 0.0906income5 0.1739** 0.0237 -0.0810** 0.0242income6 0.0456 0.0082 -0.0367** 0.0286DYYP -0.0028** 0.0001 0.0011° -0.0014unemployed -0.0680* 0.0003 -0.1683** 0.0426selfempl 0.0025 0.0035 0.0077 0.0020age < 30 0.0091 -0.0006 -0.0062° -0.0024 30/39 -0.0093** 0.0002 -0.0041° -0.0075 40/49 -0.0052* -0.0009* -0.0075** 0.0001 50/64 -0.0094** -0.0018** -0.0117** -0.0038** 65+ -0.0089** -0.0023** -0.0194** -0.0100**kids <=6yrs -0.0010 0.0036 -0.0063 0.0091kids 7-12 0.0073 0.0066** -0.0057 0.0188kids 13-19 -0.0044 0.0067** 0.0134 0.0145kids 20+ 0.0108 0.0062** 0.0102 0.0237**single -0.0822** -0.0051 -0.0875** -0.0377*===========================================================
od31
Prob(Rejection | Application) – marg. effects============================================== NL IT US ----------------------------------------------wealth -0.0002** -0.0001 -0.0089**income1 0.0035 -0.0074 0.0093 income2 -0.0033 -0.0497 -0.0832* income3 -0.0082 0.0073 -0.1585**income4 -0.0061 -0.0445 -0.2362**income5 0.0018 0.0008 -0.0975**income6 -0.0088 -0.0375 -0.0117 DYYP 0.0001 0.0015* 0.0058**unemployed 0.0028 0.0468 -0.0222 selfempl 0.0064° 0.0013 0.0241° age < 30 0.0012° 0.0015 -0.0033 30/39 -0.0002 0.0004 -0.0093** 40/49 0.0003 0.0006 -0.0006 50/64 -0.0004° 0.0003 0.0002 65+ 0.0001 0.0009 -0.0069* kids <=6yrs 0.0003 0.0049 0.0142 kids 7-12 -0.0002 0.0095 0.0354**kids 13-19 -0.0001 0.0029 0.0420**kids 20+ 0.0013 0.0036 -0.0152 single -0.0006 0.1114* -0.0086 ==============================================
od32
Prob{(Rejection | Application) or (Discouraged | No Application)} marg. effects=========================================================== NL IT US SP -----------------------------------------------------------wealth -0.0001** -0.0011** -0.0080** -0.0037**income1 0.0004 -0.0011 0.0042 -0.0081income2 -0.0013 -0.0026 -0.0158 -0.0159income3 -0.0032 -0.0127 -0.1227** -0.0005income4 -0.0001 -0.0093 -0.1585** -0.0004income5 0.0030 0.0057 -0.0780** -0.0227income6 -0.0017 -0.0077 -0.0150 -0.0069DYYP 0.0000 0.0003** 0.0040** -0.0003unemployed 0.0006 0.0160** -0.0238 0.0216*selfempl 0.0028** 0.0017 0.0182 -0.0006age < 30 0.0001 -0.0002 -0.0031 -0.0016 30/39 -0.0001* 0.0003 -0.0074** 0.0001 40/49 0.0000 -0.0006* -0.0017 0.0001 50/64 -0.0001 -0.0002 -0.0032** 0.0001 65+ -0.0001 -0.0008** -0.0115** -0.0003kids <=6yrs -0.0001 0.0026 0.0070 0.0033kids 7-12 0.0001 0.0035** 0.0262** 0.0019kids 13-19 0.0001 0.0021 0.0346** -0.0048kids 20+ 0.0000 0.0027** -0.0020 0.0020single -0.0006* 0.0030 -0.0186 -0.0022========================================================== od33
E(Debt | Debt> 0) - coefficients=================================================================== NL IT US SelCorr SP(Tobit)-------------------------------------------------------------------wealth -0.024** -0.033** -0.008* 0.010 -0.125 income1 -0.088 -0.113* -0.080** -0.075 0.291income2 0.829** 0.718** 1.940** 1.988** 5.658**income3 2.524** 0.362 1.372** 1.427** 4.176°income4 0.369 0.332 1.221** 1.219** 0.840income5 1.259** 0.306 0.973** 0.967** 2.324income6 0.175 0.651** 0.571** 0.521** -0.191DYYP -0.014** 0.001 -0.023** -0.021** 0.039unemployed -0.149 -0.058 -0.073 -0.109 0.136selfempl 0.141° 0.657** 0.416** 0.382** -0.336age < 30 0.192** -0.010 0.051** 0.042** 0.215 30/39 0.030** 0.014 0.006 0.008 -0.202* 40/49 -0.008 -0.023** -0.007 -0.014* -0.198* 50/64 -0.036** -0.017* -0.008° -0.008 -0.327** 65+ -0.031** -0.011 -0.055** -0.066** -0.459**kids <=6yrs 0.120** -0.021 0.053* 0.063* 1.661**kids 7-12 0.063° 0.026 0.047° 0.061** 0.974*kids 13-19 -0.005 -0.040 0.069** 0.061* -0.137kids 20+ -0.051 0.026 0.063° 0.047 0.891**single -0.670** -0.353** -0.138* -0.114 -3.642**=================================================================
od34
Credit behavior differs across countries:
• Much greater percentage apply for credit in US than in NL or Spain, Italy least. (Italy consistent with greater social insurance).
• Of those who apply a much higher percentage are rejected in the US than Italy or the NL. The percentage in Spain is tiny. (Consistent with more credit bureau data available in NL and US)
• Percentage who are rejected or discouraged much larger in the US, about the same in NL, Italy and Spain.
Re Application
• Less wealthy more likely to apply in all countries• Prob of application follows life cycle model wrt age• Unemployed less likely to apply in US & NL (low insurance & high insurance respectively)• Effect of permanent income consistent with PIH only for NL.
Conclusions
od35
Re Credit Constraints
• Less wealthy rejected or discouraged in all 4 countries• Income above permanent income increases chance of rejection or of being discouraged in Italy & US• Unemployed more likely to be rejected or discouraged in Italy and Spain• Self employed more likely to be rejected or discouraged in NL
Re Debt Outstanding
• Income above permanent income reduces debt outstanding in NL and US • Debt outstanding follows simulated precautionary savings model• Self employed increases debt outstanding in Italy and US (consistent with less social insurance)
od36
Combined
Difference between income and permanent income
• reduces both the chance of application and the volume for those who demand debt in NL
• has no effect on the chance of application or on the volume in IT or SP
• increases the chance of application but reduces the volume in the US
od37