![Page 1: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/1.jpg)
The supply side of housing finance G. Foa', L. Gambacorta, L. Guiso, P. Mistrulli
Yale, BIS, EIEF, Bank of Italy
NYF May 20, 2015
![Page 2: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/2.jpg)
Financial Advice • Households rely heavily on advice from financial
intermediaries: 80% in Germany, 91% UK, 73% US • Potential for biased recommendations
1. Limited information 2. Limited sophistication
• Research questions: 1. How to identify biased advice? 2. Is it quantitatively important?
• Look at mortgage type choice – ARM/FRM (temptation to bias and its cost can be large)
![Page 3: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/3.jpg)
A Test for Financial Advice • Under no advice, prices summarize supply effects on
choice • Under advice, supply shocks and incentives matter • Mortgage choice (ARM vs FRM) and prices for 2
million Italian mortgages originate (04-10) 1. Observe bank identity 2. Measure time-varying supply factors
• Bank Bond spread, core deposits, access to securitization
• Disentangle the price channel and the advice channel
![Page 4: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/4.jpg)
Approaches
• Current approaches: 1. Compare performance of advised vs non-advised
– Selection bias 2. Randomized field experiments
– External validity + long term customers
• Require to observe advice – Unsolicited
• Our approach: – no need to observe advice once we observe customers
choices, prices and banks supply factors (incentives) – Identifying assumptions needed
![Page 5: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/5.jpg)
Results • We detect strong supplier effects=> consistent
with biased advice • Relative price of ARM/FRM main determinant of
choice, but supply factors matter – E.g. borrowers from banks that experienced an
increase in bond spread, more likely to take ARM – Bank bond spread effect is 10% of FRM/ARM price
• Additionally 1. Results stronger under price inaction 2. Effects stronger for unsophisticated households
![Page 6: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/6.jpg)
Outline
• Theory
• Empirical strategy
• Data
• Results
![Page 7: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/7.jpg)
Mortgage Choice
• Households – Get a mortgage. Choice FRM versus ARM – Risk: income, inflation, real rate – is the FRM premium – G is the distribution of risk aversion
• Koijen rule
is the ARM share. Choice is
![Page 8: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/8.jpg)
Banks
• Continuum of regions, distribution G in each region – Banks are local monopolies – Banks get (heterogeneous) supply shocks – what is ? Access to LTF, deposit base, securitization
• Payoff of the bank – x = is the ARM share
![Page 9: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/9.jpg)
No Advice (all weary)
• Bank’s problem
• Optimization gives so that decision rule (simplified) is Proposition 1: under no advice mortgage choice depends on supply factors only through prices
![Page 10: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/10.jpg)
Distorted Advice
![Page 11: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/11.jpg)
Distorted Advice: FOCs • Bank payoff
• FOCs Naïve =>
![Page 12: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/12.jpg)
Advice Proposition 2: under advice (if # of )
• FOCs are a joint restriction on • But unobservable
• add information (unless are linked by a
deterministic relationship - always true if )
![Page 13: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/13.jpg)
Costly price adjustment Assume menu cost to change => price inaction Proposition 3: under advice and price inaction 1. Even if # of
2. Effect of on mortgage choice stronger during inaction
![Page 14: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/14.jpg)
Takeaways • Under no advice, supply factors do not affect mortgage
choice conditional on prices • Under advice they do
1. We test for 2. Signs should go in the “right” way:
• Under price rigidity the result is always true
![Page 15: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/15.jpg)
Main Equation
![Page 16: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/16.jpg)
Identification • Source of failure: sorting
– e.g. Larger banks attract more risk averse customers
• Include a bank fixed effects together with time varying supply factors – Key assumption: the composition of the pool of borrowers
does not react to bank specific quarter to quarter variation in funding conditions
• Sorting unlikely to be an issue in our data
![Page 17: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/17.jpg)
Price inaction
![Page 18: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/18.jpg)
Data Sources • Bank of Italy datasets:
– Credit Register: info on loans exposure (above 75k) – SLIR: survey data on interest rate charged on loans (175 banks) – Banks broad geographic coverage (median: 80% of provinces)
• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln
• Relevant info: – Mortgage info: amount, rate, type – Borrower info: age, gender, nationality, province, cohabitation,
distance – Lender info: identifier => balance sheet information
descr stats
![Page 19: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/19.jpg)
The Relative Price • Relative price
• Problem: do not observe both for single i • Impute them for each b from
• Compute FRM risk premium (1 year lag mov av)
• Issue: some measurement error in link
![Page 20: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/20.jpg)
Supply Factors
• Bank bond spread=> relative advantage in ARM • Securitization activity => relative advantage in
FRM (Fuster & Vickery, 2014)
• Deposit to total funding => relative advantage in FRM (Berlin & Mester, 1999)
![Page 21: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/21.jpg)
Price Inaction
• Compute for each b,t
• Define inaction =1 if
• Also with Sd/4 link
![Page 22: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/22.jpg)
Price Inaction: evidence 0
.51
1.5
2
-3 -2 -1 0 1 2 3
Density kdensity change_spread
All sample : 2004 - 2010
Distribution of the changes of the spread
0.1
.2.3
.4fr
equ
en
cy
0 2 4 6 8 10 12 14 16Quarters bank remained inactive
Density kdensity counter
Cross sectional distribution of the numbers of quarters banks remained inactive (28 Q): median 11
Price inaction in about 50% of the observations
![Page 23: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/23.jpg)
ARM Share
![Page 24: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/24.jpg)
Lender Characteristics & Mortgage Choice lhs=1 if FRM chosen
1 percentage point increase (1 sd) in the FRM premium lowers the fraction of FRM by 34 pp
I II III IV
Long Term Financial Premium (LTFP) -0.31*** -0.35*** -0.34***
(0.029) (0.027) (0.026) Bank fixed effects yes yes yes yes
Time fixed effects no no yes yes
Borrowers' Characteristics no no no yes
Province fixed effects no no no yes
Observations 1,662,429 1,662,429 1,662,429 1,662,429
R-squared 0.098 0.476 0.592 0.600
![Page 25: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/25.jpg)
Supply effects
RC
Baseline model non- linear terms for LTP
LTFP -0.349*** -0.474***
LTFP2 -0.012
LTFP3 0.028***
Bank bond spread -0.027* -0.029*
(0.015) (0.017) Securitization activity 0.138*** 0.124***
(0.028) (0.023)
Deposit ratio % 0.006*** 0.006***
(0.002) (0.002)
Bank fixed effects yes yes
Time fixed effects yes yes
Borrowers' Characteristics yes yes Province fixed effects and control for bank competition yes yes
Observations 1,662,389 1,662,389
R-squared 0.608 0.628 RC
![Page 26: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/26.jpg)
Static Sorting? link
![Page 27: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/27.jpg)
“Dynamic” Sorting? Mortgage size
(log) Italian Cohabitation Age Female
Bank bond spread -0.0010 0.0007 0.0009 -0.0251 -0.0016
(0.0067) (0.0045) (0.0022) (0.0775) (0.0012)
Deposit ratio -0.0001 -0.0001 0.0001 -0.0047 -0.0000
(0.0005) (0.0006) (0.0004) (0.0104) (0.0001) Securitization activity -0.0172 0.0036 -0.0103 -0.4767 0.0024
(0.0247) (0.0191) (0.0090) (0.3757) (0.0033) Bank fixed effects yes yes yes yes yes
Time effects yes yes yes yes yes Province fixed effects yes yes yes yes yes
Observations 1,600,309 1,600,309 1,600,309 1,600,309 1,600,309
R-squared 0.0413 0.0613 0.0179 0.0347 0.0030
![Page 28: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/28.jpg)
Yet, Individual Characteristics Affect Contract Choice
Log Mortgage size -0.043*** (0.007)
Female 0.011*** (0.001)
Age -0.0004* (0.0002)
Italian 0.049*** (0.008)
Joint mortgage 0.007** (0.003)
Cohabitation -0.003** (0.0014)
BFE, TFE, PFE, Bank competition and other controls yes Observations 1,662,389 Adjusted R-squared 0.360
![Page 29: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/29.jpg)
Demand Shocks Driving Results?
Controlling for Time- Province FE
Only banks present in all provinces
LTFP -0.280*** -0.404*** (0.021) (0.034)
Bank bond spread (2) -0.027* -0.026*
(0.015) (0.015) Securitization activity (3) 0.132*** 0.223***
(0.030) (0.04) Deposit ratio % (4) 0.005*** 0.009***
(0.001) (0.002) Observations 1,662,389 957,961 Adjusted R-squared 0.5729 0.6615
![Page 30: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/30.jpg)
Effects Under Price Inaction Baseline
LTFP (1) -0.350*** (0.024)
Bank bond spread -0.014 (0.016)
Securitization activity 0.137*** (0.025)
Deposit ratio % 0.005*** (0.002)
Dib 0.052* (0.031)
Bank bond spr * Dib -0.0621*** (0.013)
Secur- Activity * Dib 0.017* (0.010)
Dep ratio * Dib 0.0008* (0.0005)
Bank, Time, Province FE , Borrowers' Characteristics yes Observations 1,662,389 R-squared 0.609
![Page 31: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/31.jpg)
Effect of Sophistication link (a)
Sophisticated borrowers:
(b) Unsophisticated
borrowers:
Difference
|b-a|
H0: |b-a|>0
Long Term Financial Premium (LTFP) (1) -0.315*** -0.397*** 0.082 **
(0.025) (0.029) (0.039)
Bank bond spread (2) -0.013 0.007 0.021
(0.019) (0.024) (0.030)
Securitization activity (3) 0.108*** 0.175*** 0.066 **
(0.024) (0.019) (0.031)
Deposit ratio % (4) 0.005*** 0.007*** 0.002 *
(0.001) (0.001) (0.001)
Dib (5) 0.060 0.046 0.014
(0.039) (0.028) (0.048)
Bank bond spread * Dib -0.036** -0.085*** 0.048 **
(0.015) (0.025) (0.029)
Securitization Activity * Dib -0.017 0.027* 0.045 **
(0.021) (0.015) (0.026)
Deposit ratio % * Dib -0.001 0.001** 0.003 **
(0.002) (0.0004) (0.002) (BFE, TFE, PFE), Borrowers characteristics yes yes Observations 29 527 27 158
![Page 32: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/32.jpg)
Alternative explanation?
• Advertisement, not distorted advice • Banks target share of FRM/ARM: say 70/30
and then rationing – Rationing stronger at times of price inaction – Stronger for unsophisticated if have higher search
costs (take the contract offered rather than move)
• Both imply sorting/selection • We do not see it in the data
![Page 33: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/33.jpg)
Summarizing • Evidence consistent with biased advice
• Quantitatively important. 1 sd QoQ increase in:
– bond spread ) +2.8 pp in Pr(ARM) – entry in sec mkts ): – 3.1 pp in Pr(ARM) – Deposits/Funding ): -3.2% in Pr(ARM)
• Inaction and sophistication reinforce results
![Page 34: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/34.jpg)
Next Steps • Deal with sorting formally • Observe banks visited (banks have to send an inquire
to the Credit Register) • Two stage (Heckman type) model:
– Bank selection: driven by level of interest rates – Mortgage type: driven by FRM premium
• Same FRM spread consistent with different levels of interest rates (natural exclusion restrictions)
• Extend model to deal with competition – Price (comparable) versus advice (harder to compare)
under competition
![Page 35: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/35.jpg)
Measurement error: link
• Imputation may be problematic: – Measurement error => Φ endogenous – Φ and B correlated β3 may be biased
• We restrict analysis to inaction times – Φ and B uncorrelated in inaction times
• Inaction orthogonal to selection within the bank
![Page 36: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/36.jpg)
Supply shocks orthogonal to prices during inaction back
![Page 37: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/37.jpg)
Bank fixed effect distribution back 0
12
34
frequ
ency
-1.2 -1 -.8 -.6 -.4 -.2 0 .2Fixed effects
Density kdensity EF
Bank specialization
![Page 38: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/38.jpg)
Rationing back
• Banks target share of FRM/ARM: say 70/30 • They then ration contract types
– Rationing stronger at times of price inaction – Stronger for unsophisticated if have higher search
costs (take the contract offered rather than move) • But rationing=> sorting which we do not see
– Bank 70 FRM/30 ARM=> should attract disproportionate share of risk averse compared to bank targeting 30 FRM/70 ARM
– We do not see it in the data
![Page 39: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/39.jpg)
Reverse causality back
• Two answers: 1. unlikely as it takes time to securitize or to
attract deposits 2. Cannot explain sign of Bank Bond Spread
1. Increase in demand of FRM=> Issue more fixed rate bonds => pay higher spreads => positive correlation between FRM choice and Bank Bond Spread
2. Correlation is negative
![Page 40: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/40.jpg)
Financial Sophistication back
• H0: Biased advice less important for sophisticated borrowers
• Run the experiment for sophisticated/unsophisticated borrowers and test for difference
• Proxies: 1. Value of the mortgage ) proxies for wealth
(education) 2. Have already borrowed in the past vs first time
borrowers
![Page 41: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/41.jpg)
Descriptive stats: 1 Mortgage characteristics
![Page 42: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/42.jpg)
Descriptive stats:2
Individual characteristics
![Page 43: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/43.jpg)
Descriptive stats:3 back
Lender characteristics
![Page 44: The supply side of household finance...• Data: 2 mln mortgages 04-10. Focus on comparable contracts: standard 20-25 year FRM and ARM – End up with 80% of the sample 1.6 mln •](https://reader034.vdocument.in/reader034/viewer/2022051822/5fec069ec4c5fd21ec72815b/html5/thumbnails/44.jpg)
Descriptive stats:3
Stats for the time varying component of supply factors