Household finance. James J. Choi October 13, 2017

Transcription

1 Household finance James J. Choi October 13, 2017

2 Household finance How households use financial instruments to attain their objectives --John Campbell AFA Presidential address, 2006 Saving Asset allocation Spending out of savings Borrowing 2

3 Propellants of growth New data Big Data Great Recession: Household finance matters for aggregates! 3

4 Outline Savings adequacy Retirement expenses Pre-retirement expenses Financial literacy Savings nudges Household capital structure 4

5 Many Americans arrive at retirement with no liquid wealth Net worth excluding pensions, student loans, durables, homes, and collateralized debts, ages $90,000 $80,000 $70,000 $60,000 $50,000 $40,000 $30,000 $20,000 $10,000 $0 $77,184 $1 $6,213 25th percentile 50th percentile 75th percentile Source: 2013 Survey of Consumer Finances 5

6 Adding defined contribution pensions doesn t really affect the left tail $300,000 $250,000 $200,000 $150,000 Net worth excluding defined benefit pensions, durables, homes, and collateralized debts, ages $265,905 $100,000 $50,000 $0 $305 $41,776 25th percentile 50th percentile 75th percentile Source: 2013 Survey of Consumer Finances 6

7 The left tail accumulates wealth mainly through illiquid home equity $800,000 $700,000 $600,000 $500,000 $400,000 $300,000 $200,000 $100,000 $0 Net worth excluding defined benefit pensions, ages $50,885 $220,847 $672,557 25th percentile 50th percentile 75th percentile Source: 2013 Survey of Consumer Finances 7

8 Are American undersaving? Scholz, Seshadri, and Khitatrakun (2006) Build lifecycle savings model Find that 84% of age households in 1992 Health and Retirement Study are at or above optimal savings level Median deficit among those below target is $5,260 But left tail s net worth is mostly housing equity Most households don t use housing equity to finance non-housing consumption in early decades of retirement (Poterba, Venti, and Wise, 2011) 8

9 Consumption drop at retirement Aguiar and Hurst (2005) Food expenditure drops but not food consumption (or quality) on average across retirement threshold BUT among retirees with < $1,000 in liquid assets and no home ownership (bottom wealth decile), 19% decline in calories consumed Hurst (2008): Lifecycle model has a hard time matching the magnitudes of the decline in expenditures for households in the bottom quartile of the wealth distribution 9

10 Expenditure paths in retirement Single females Expenditure (age 65 = 100) college grad high school grad < high school Source: Hurd and Rohwedder (2012) Age 10

11 Most people don t want decreasing income in retirement 60% 50% Preferences over annuity income growth paths 50% 40% 30% 20% 10% 19% 32% 0% -2% real growth 0% real growth 2% real growth Source: Beshears, Choi, Laibson, Madrian, and Zeldes (2014) 11

12 Dying with no assets Sample: Households whose head was age 70+ in 1993 In the last two years before death, 40% had < $20,000 of annuity income and < $10,000 of financial assets Of these 40%, 55% also had zero home equity Poterba, Venti, and Wise (2012) 12

13 Financial fragility 46% of American adults say they could not come up with $400 to cover an emergency without borrowing or selling something (Board of Governors, 2016) 13

14 Most hand-to-mouth households have illiquid assets Kaplan, Violante, and Weidner (2014) 14

15 Why live hand-to-mouth with illiquid assets? Kaplan and Violante (2014) Illiquid assets earn illiquidity premium So worthwhile to invest all wealth in illiquid assets and suffer welfare losses from unsmoothed consumption Angeletos et al. (2001) Households have self-control problems and know it Invest in illiquid assets in order to restrain spending 15

16 Is hand-to-mouth optimal choice? Social Security benefits paid on 2nd, 3rd, or 4th Wednesday of each month Based on day of the month you were born Four months per year have five Wednesdays Causes pay cycles to be 35 days instead of 28 Baugh, Leary, and Wang (2017) 16

17 Within-month financial distress Compared to 2nd Wednesday group, 4th Wednesday group is 3% less likely to experience overdraft 10% less likely to bounce a check 14% less likely to get online payday loan 4% less likely to get storefront payday loan In 35-day pay period 5% more likely to experience overdraft 3% more likely to bounce a check 16% more likely to get online payday loan 31% more likely to get storefront payday loan Baugh, Leary, and Wang (2017) 17

18 Measuring financial literacy: The Big Three questions Suppose you had $100 in a savings account and the interest rate was 2% per year. After 5 years, how much do you think you would have in the account if you left the money to grow? More than $102 Exactly $102 Less than $102 Imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. After 1 year, would you be able to buy more than today, exactly the same as today, or less than today with the money in this account? More than today Exactly the same today Less than today Source: Lusardi and Mitchell (2008) 18

19 Measuring financial literacy: The Big Three questions Do you think that the following statement is true or false: buying a single company stock usually provides a safer return than a stock mutual fund? True False Source: Lusardi and Mitchell (2006) 19

20 Measured financial literacy Interest USA Netherlands Japan Germany Chile Mexico Correct 78% 85% 71% 82% 47% 45% Don t know 10% 9% 13% 11% 32% 2% Inflation Correct 65% 77% 59% 78% 18% 71% Don t know 19% 14% 29% 17% 21% 2% Risk Correct 53% 52% 40% 62% 41% 47% Don t know 40% 33% 56% 32% 33% 1% All correct 39% 45% 27% 53% 8% 15% Source: Hastings, Madrian, and Skimmyhorn (2013) 20

21 Does more extensive financial education help? Expansion over time in state high school financial education mandates Bernheim, Garrett, and Maki (2001): Positive wealth accumulation effect Cole and Shastry (2010): No effect Mandatory 8-hour financial literacy course (plus assistance in enrolling in Thrift Savings Plan) for new soldiers in U.S. Army, roll-out staggered across bases (Skimmyhorn, 2016) Avg. contribution to TSP is $20/month higher than control in first year after, $14/month higher than control in second year Debt balance is $608 lower than control in first year after, $202 lower than control in second year 21

22 Do people get professional advice? 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% % using financial advisor < High school High school College+ Lusardi, Michaud, and Mitchell (2017) 22

23 One solution: Forced savings Singapore Compulsory saving of 37% of covered wages until age 50, lower percent after that 20% employee contribution 17% employer contribution 23

24 24

25 The old routine when you joined a company with a 401(k) plan Welcome to the company Here is information on your 401(k) plan If you d like to join, call this toll-free number or visit the benefits website 25

26 Automatic 401(k) enrollment Welcome to the company Here is information on your 401(k) plan If you don t do anything before a deadline, you will be automatically enrolled at this default contribution rate and asset allocation If you d like to opt out, call this toll-free number or visit the benefits website 26

27 Automatic enrollment effect Fraction enrolled in 401(k) 100% 80% 60% 40% 20% 0% Tenure (months) Hired and observed before automatic enrollment Hired under automatic enrollment (3% contribution default) Hired under automatic enrollment (6% contribution default) Source: Beshears, Choi, Laibson, and Madrian (2008) 27

28 Autoenrollment in U.S. Legislatively encouraged by Pension Protection Act of % of 401(k) plans in 2015 automatically enrolled employees (Plan Sponsor Council of America, 2016) 28

29 Autoenrollment in UK Mandatory automatic enrollment being phased in from by firm size Chancellor George Osborne: Biggest changes to pensions in 100 years To date, opt-out rate of only 9-10% a surprising shock, with the DWP initially expecting a 28% opt-out ( 29

30 Aggregate effect of UK autoenrollment Source: Automatic enrolment, Commentary and analysis: April 2015 March The Pensions Regulator. automatic-enrolment-commentary-analysis-2016.pdf 30

31 Defaults and herding Fraction of Participants % 3% 4-5% 6% 7-10% 11-15% Contribution Rate Hired before automatic enrollment Hired during automatic enrollment (3% default) Hired during automatic enrollment (6% default) Choi, Laibson, Madrian, and Metrick (2006) 31

32 Defaults and herding Before Automatic Enrollment Money Market 7% Bonds 18% After Automatic Enrollment Money Market Default Fund Stocks 16% Bonds 3% Madrian and Shea (2001) 75% Stocks 81% Money Market 32

33 How sticky are defaults? 100% Fraction of participants 80% 60% 40% 20% 0% Tenure (months) Hired before AE: Default rate and fund Hired after AE: Default rate and fund Hired before AE: Default rate (3%) Hired after AE: Default rate (3%) Hired before AE: 100% in default fund Hired after AE: 100% in default fund Choi, Laibson, Madrian, and Metrick (2004) 33

34 How sticky are defaults? Fraction of participants 80% 60% 40% 20% 0% Tenure (months) Hired before AE: Default rate and fund Hired after AE: Default rate and fund Hired before AE: Default rate (2%) Hired after AE: Default rate (2%) Hired before AE: 100% in default fund Hired after AE: 100% in default fund Choi, Laibson, Madrian, and Metrick (2004) 34

35 Opt-out delay from a 0% contribution default Mean time to enrollment (days) Contribution rate chosen upon enrollment (% of salary) Carroll, Choi, Laibson, Madrian, and Metrick (2009). Area of bubble is proportional to number of employees at that contribution rate. 35

36 Dynamic defaults: Save More Tomorrow 401(k) contribution rate rises automatically in the future Rise may coincide with pay raises Thaler and Benartzi (2004) 36

37 Auto-escalation effect Contribution rate 16% 14% 12% 10% 8% 6% 4% 2% 0% Before Pay Pay Pay Pay raise 1 raise 2 raise 3 raise 4 Accepted advice to save more now Joined SMT Thaler and Benartzi (2004) 37

38 Auto-escalation works better if it s the default SMT participation rate 100% 80% 60% 40% 20% 0% -Q4 -Q3 -Q2 -Q1 +Q1 +Q2 +Q3 +Q4 Time relative to auto-enroll into SMT Benartzi, Peleg, and Thaler (2012) 38

39 Why do defaults work? Some candidate mechanisms Opting out requires paying effort cost Exacerbated by time inconsistency Belief that default is a recommended choice Unawareness that default exists or can opt out of it Cognitive dissonance causes people who find themselves at default to manufacture reasons why it s the right choice Default serves as an anchor Individuals consider only a subset of possible choices, and default is disproportionately likely to be in that subset Default becomes a reference point around which gains and losses are evaluated 39

40 Active choice Welcome to the company You have 30 days to tell us whether you want to be in the 401(k) plan Not stating a preference is not an option Carroll, Choi, Laibson, Madrian, and Metrick (2009) 40

41 Active choice 401(k) participation effect Fraction enrolled in 401(k) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Tenure at company (months) Active decision cohort Standard enrollment cohort Carroll, Choi, Laibson, Madrian, and Metrick (2009) 41

42 Active choice 401(k) contribution rate effect Average 401(k) contribution rate (non-participants included) 6% 5% 4% 3% 2% 1% 0% Tenure at company (months) Active decision cohort Carroll, Choi, Laibson, Madrian, and Metrick (2009) Standard enrollment cohort 42

43 Active choice vs. asymptotic opt-in contribution rate distributions 43

44 Active choice vs. asymptotic opt-in contribution rate distributions 44

45 Do savings nudges increase total savings? Chetty et al. (2014) Danish data that include measure of total wealth Policy studied: Changes in mandatory employer pension contributions when people switch jobs Not automatic enrollment, and not a nudge But can be undone by people not at a corner 45

46 Do savings nudges increase total savings? Chetty, Friedman, Leth-Petersen, Nielsen, and Olsen (2014) 46

47 How long does the savings increase last? Chetty, Friedman, Leth-Petersen, Nielsen, and Olsen (2014) 47

48 Debt overhang (Melzer, 2017) Idea For homeowners with negative equity, home improvements benefit creditor, not homeowner Therefore, invest less in home Main data Consumer Expenditure Survey 48

49 Home improvements Loan-to-value ratio (%) 49

50 Unscheduled principal payments Loan-to-value ratio (%) 50

51 Debt overhang results Homeowners with negative equity Spend $200 (30%) less per quarter on home improvements and maintenance Cut unscheduled principal payments by $180 (40%) Differences not explained by income, total expenditures, financial wealth, demographics, property characteristics, mortgage traits Estimate that debt overhang reduced national spending on home improvements by 3-5% per year, >10% in Arizona, California, Florida, and Nevada 51

52 Empirical challenge Negative equity homeowners may be more financially constrained Arguments against this confound Results also hold for high-income households with substantial financial assets Results weaker in recourse states Spending on appliances, furniture, entertainment durables, jewelry, and vehicles unaffected by negative equity after controls Holds when comparing two properties, one underwater and the other not, owned by a single person 52

53 Debt overhang (Bernstein, 2016) Idea Income-contingent distressed mortgage modification acts as an implicit tax on labor supply More earnings more repayments to creditor Effect: Reduce earnings Main data Transaction-level bank/credit card/mortgage account data from financial institution covering >25% of U.S. households, Restriction to households with main bank account and mortgage at data provider ~200,000 households 53

54 Debt overhang (Bernstein, 2016) Empirical challenge Economic distress causes both negative equity and reduced job opportunities Instrument while controlling for region time fixed effects Loan-to-value if house appreciation since mortgage = regional avg. rate and repayment rate was minimum under 30-year fixed rate with national median mortgage rate Variation driven by when moved to region 54

55 Effect of negative equity LTV > % decline in income Effect stronger in regions with higher mortgage modification rate, controlling for delinquency rate Effect stronger in states with judicial foreclosure requirements Harder to foreclose, so more modifications 55

56 Payment priority (Gathergood, Mahoney, Stewart, and Weber, 2017) How do individuals choose how much to pay back on each credit card? Optimal behavior with two cards Pay minimum on each card Any extra payment goes to highest interest card Only pay lower interest card if other card paid off in full Data: 1.4 million individuals in U.K across five major card issuers,

57 Does this choice matter? Among those who hold exactly 2 cards Average APR difference between cards: 6.5 percentage points Average APR level: 19.7% Average revolving balance on higher-apr card: 2,198 Average revolving balance on lower-apr card: 2,049 57

58 Repayment behavior, 2 cards 58

59 Repayment behavior, 3 cards 59

60 Repayment behavior, 4 cards 60

61 Repayment behavior, 5 cards 61

62 Rational inattention? 62

63 Rational inattention? 63

64 Inexperience? 64

65 What s going on? Balance matching Match share of payments to share of balances on card Consistent with heuristic such as pay 10% of each card s balance 65

66 Bankruptcy policy (Yannelis, 2016) 11.5% of federal student loan borrowers who began repayment in Oct defaulted by Sep Should we allow these defaulters to discharge their student debt in bankruptcy? Answer partially depends on how much default is strategic 66

67 Two reforms Bankruptcy discharge reform Before 1998, could discharge student loans in bankruptcy after 7 years in repayment Starting in 1998, student loans almost completely nondischargeable Wage garnishment reform Before 2006, defaulted student loan borrowers subject to wage garnishment of 10% above threshold Starting in 2006, garnishment rate increased to 15% Both reforms do not affect current liquidity of borrowers 67

68 Data National Student Loan Data System Contains all federally guaranteed student loans (92% of all student loans in ) IRS data from Compliance Data Warehouse W-2 forms Schedule C (business income) 68

69 Defaults by repayment year, vs cohorts 69

70 Diff-in-diffs of defaults across garnishment threshold following wage garnishment reform 70

71 Results summary Student loan borrowers who can discharge student debt in bankruptcy are 18% more likely to default When garnishment rate increases by 50%, additional $10,000 of garnishable income leads to 15% decrease in default rate 71

72 Summary Most Americans arrive at retirement with almost no liquid wealth Result: Financial fragility during working life Evidence that at least the left tail of Americans doesn t save enough for retirement Financial literacy is low Financial education has only modest effects 72

73 Summary Nudges like automatic enrollment and active choice can increase retirement account balances Some evidence that increases total savings, not just reshuffling of assets Lots of opportunity to study household capital structure 73

74 Aguiar, Mark, and Erik Hurst, Consumption versus expenditure. Journal of Political Economy. Angeletos, George-Marios, David Laibson, Andrea Repetto, Jeremy Tobacman, and Stephen Weinberg, The hyperbolic consumption model: Calibration, simulation, and empirical evaluation. Journal of Economic Perspectives. Baugh, Brian, Jessse Leary, and Jialan Wang, When is it hard to make ends meet? University of Illinois working paper. Benartzi, Shlomo, Ehud Peleg, and Richard H. Thaler, Choice architecture and retirement savings plans. In The Behavioral Foundations of Public Policy, Eldar Shafir, editor. Bernheim, B. Douglas, Daniel Garrett, and Dean Maki, Education and saving: The longterm effects of high school financial curriculum mandates. Journal of Public Economics. Bernstein, Asaf, Negative equity, household debt overhang, and labor supply. University of Colorado working paper. Beshears, John, James Choi, David Laibson, and Brigitte C. Madrian, The Importance of Default Options for Retirement Saving Outcomes: Evidence from the United States. In Lessons from Pension Reform in the Americas, Stephen J. Kay and Tapen Sinha, editors. John Beshears, James Choi, David Laibson, Brigitte Madrian, and Stephen Zeldes, What makes annuitization more appealing? Journal of Public Economics. Board of Governors of the Federal Reserve System, Report on the Economic Well-Being of U.S. Households in Carroll, Gabriel D., Choi, James J., David Laibson, Brigitte C. Madrian, and Andrew Metrick, Optimal Defaults and Active Decisions. Quarterly Journal of Economics. Chetty, Raj, John Friedman, Soren Leth-Petersen, Torben Nielsen, and Tore Olsen, Active vs. passive decisions and crowd-out in retirement savings accounts: Evidence from Denmark. Quarterly Journal of Economics. Choi, James J., David Laibson, Brigitte C. Madrian, and Andrew Metrick, For Better or For Worse: Default Effects and 401(k) Savings Behavior. In Perspectives in the Economics of Aging, David A. Wise, editor. Choi, James J., David Laibson, Brigitte C. Madrian, and Andrew Metrick, Saving for Retirement on the Path of Least Resistance. In Behavioral Public Finance: Toward a New Agenda, Ed McCaffrey and Joel Slemrod, editors.

75 Cole, Shawn, and Gauri Shastry, Is High School the Right Time to Teach Savings Behavior? The Effect of Financial Education and Mathematics Courses on Savings. Working paper. Gathergood, John, Neale Mahoney, Neil Stewart, and Jörg Weber, How do individuals repay their debt? The balance-matching heuristic. University of Nottingham working paper. Hastings, Justine, Brigitte Madrian, and William Skimmyhorn, Financial Literacy, Financial Education, and Economic Outcomes. Annual Review of Economics. Michael Hurd and Susann Rohwedder, Economic preparation for retirement. In Investigations in the Economics of Aging, David A. Wise, editor. University of Chicago Press. Hurst, Erik, Understanding consumption in retirement: Recent developments. In Recalibrating Retirement Spending and Savings, John Ameriks and Olivia Mitchell, editors. Oxford University Press. Kaplan, Greg, and Giovanni Violante, A model of the consumption response to fiscal stimulus payments. Econometrica. Kaplan, Greg, Giovanni Violante, and Justin Weidner, The wealthy hand-to-mouth. Brookings Papers on Economic Activity. Lusardi, Annamaria, Pierre-Carl Michaud, and Olivia Mitchell, Optimal financial knowledge and wealth inequality. Journal of Political Economy. Lusardi, Annamaria, and Olivia Mitchell, Planning and Financial Literacy: How Do Women Fare? American Economic Review. Madrian, Brigitte C., and Dennis Shea, The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior. Quarterly Journal of Economics. Melzer, Brian, Mortgage debt overhang: Reduced investment by homeowners at risk of default. Journal of Finance. Plan Sponsor Council of America, th Annual Survey of Profit Sharing and 401(k) Plans. Chicago, IL: Plan Sponsor Council of America. Poterba, James, Steven Venti, and David Wise, The composition and drawdown of wealth in retirement. Journal of Economic Perspectives. Poterba, James, Steven Venti, and David Wise, Were they prepared for retirement? Financial status at advanced ages in the HRS and AHEAD Cohorts. In Investigations in the Economics of Aging, David A. Wise, editor. University of Chicago Press, pp

76 Scholz, John Karl, Ananth Seshadri, and Surachai Khitatrkaun, Are American saving optimally for retirement? Journal of Political Economy. Skimmyhorn, William, Assessing financial education: Evidence from boot camp. American Economic Journal: Economic Policy. Thaler, Richard H., and Shlomo Benartzi, Save More Tomorrow: Using Behavioral Economics to Increase Employee Savings. Journal of Political Economy. Yannelis, Constantine, Strategic default on student loans. NYU working paper.