Is a Student Loan Cii Crisis on the Horizon? Understanding Changes in the Distribution of Student t Loan Debt over Time Beth Akers, Matthew Chingos, and Alice Henriques Brown Center on Education Policy Brookings Institution Upjohn Institute Student Loans Conference University of Michigan October 26, 2013
Big Bad Student t Debt $1 trillion oh no! Outliers make good stories Typical debt burdens less sensational, but lots of evidence that they are growing Could dynamics of increasing debt and stagnant incomes create a crisis in the market for student loans?
Research Questions How has the incidence and distribution of student loan debt changed over time? How has the financial well being of student borrowers changed over time? Which student borrowers face the greatest financial hardship as a result of student debt? Today: Question #1
Data on Debt IPEDS: based on survey of institutions, only capture federal loans NCES (e.g., B&B, BPS): detailed, but selected cohorts and limited period of observations College Board: self reported Consumer Credit Panel: longitudinal, but no g, link to earnings
Survey of Consumer Finances Nationally representative cross section section of households administered every 3 years We use data from 1989 to 2010 Captures current loan balances and has multiple measures of earnings
Survey of Consumer Finances We examine total outstanding education debt of households with average age 20 40 Reflects borrowing and repayment to date Debt calculated on per person basis in 2010 dollars Apply survey weights throughout
Part I: Trends Over Time Big increases (tripling) in incidence and borrowing People are more likely to borrow and they re borrowing more
Table 1. Incidence and Amount of Debt Over Time, Age 20-40 Year Incidence Mean Debt Those with Debt Mean Median Clli Cell size 1989 14% $806 $5,810 $3,517 971 1992 20% $1,498 $7,623 $3,730 1323 1,323 1995 20% $1,475 $7,521 $3,577 1,429 1998 20% $2,539 $12,826 $8,027 1,362 2001 22% $2,881 $12,939 $6,156 1,307 2004 24% $3,402 $14,204 $7,503 1,246 2007 28% $4,583 $16,322 $9,728 1144 1,144 2010 36% $6,502 $17,916 $8,500 1,865
Trends in Debt over Time, Households with Average Age, g g 20-40, 1989-2010
Cumulative Distribution of Education Debt, Households with Average Age 20-40, 1989/1992 and 2010.6 Cumula ative Percent.7.8.9 tage 1 0 20000 40000 60000 80000 100000 Total Education Debt 1989/1992 2010
Tracking Cohort Debt over Time, Age 20-25 in 1989/1992 g, g 5 9 9/ 99 through Age 38-43 in 2007/2010
Part II: Decomposition Debt levels and trends vary by educational attainment We might worry less about increases in debt that are driven by sound investments in education
Table 2b. Summary Statistics, Individual Level, Members of Households with Average Age 20-40 Year Maximum Education of Household HS or less Some Coll BA Graduate Female Debt 1989 50% 26% 17% 6% 52% $806 1992 45% 27% 21% 7% 52% $1,502 1995 43% 30% 20% 7% 53% $1,475 1998 44% 29% 19% 8% 53% $2,539 2001 45% 26% 21% 8% 54% $2,881 2004 41% 30% 21% 9% 53% $3,402 2007 40% 31% 20% 9% 53% $4,583 2010 38% 31% 22% 9% 53% $6,502
Incidence of Debt by Educational Attainment, 1989-2010
Average Debt by Educational Attainment, Among g y, g Those with Debt, 1989-2010
Simple Reweighting Step 1: Calculate 2010 debt by educational attainment Step2: Calculateweighted average of Step 1 using 1989 educational attainment as weights 1989 average debt: $806 2010 average debt: $6,502 (increase of $5,696) 2010 reweighted: $5,343 (increase of $4,537) > explains 20 percent of increase
Reweighted (simple method, education only)
Multivariate i t Reweighting
Table 3. Summary Statistics, Household Level, Average Age 20-40 1989 2010 2010 reweighted ih Maximum Education HS Sor less 41% 31% 42% Some Coll 29% 32% 29% BA 20% 24% 20% Graduate 9% 13% 9% Race/Ethnicity of Household Head White 72% 62% 72% Black 11% 15% 11% Hispanic 11% 17% 11% Oh Other 6% 6% 6%
Multivariate i t Reweighting 1989 average debt: $806 2010 average debt: $6,502 ($5,696 increase) 2010 reweighted: $4,932 Changes in educational attainment and demographics explain 28 percent of increase
Distribution of Education Debt, Reweighting
Tuition Adjustment: t Ideal Step 1: Measure e how much each individual dua paid for their education Step 2: Measure how much they would have paid 21 years prior Step 3: Calculate the causal effect of price on debt Step 4: Calculate how much less debt they would have taken out had they faced the prices from 21 years prior (Step 3 x [Step 1 Step2])
Tuition Adjustment: t Actual Don t know how much respondents paid for education Don t know causal effect of price on debt Instead: deflate 2010 debt to simulated 1989 level using published tuition and fees by year Assumption (wrong): debt increase (due to p ( g) ( tuition) is same as percentage increase in tuition
Tuition Adjustment: t Actual Counterfactual debt in 2010 = (Actual Debt in 2010) x (average tuition 21 years prior to when respondent was age 20)/(average tuition when respondent was age 20) Tuition is enrollment weighted average of published tuition fees at two year, public fouryear, and private four year institutions Overstates importance of published tuition since increases in net price were smaller
Trends in Published Tuition and Fees, 1971-2012
Table 4. Decomposition of Changes in Mean Debt, 1989-2010 Mean Debt Change from 1989 Share of Change Explained 1989 Debt $806 2010 Debt No Adjustment $6,502 $5,696 0% Applying 1989 characteristics $4,932 $4,126 28% Applying 1989 tuition $3,194 $2,388 58% Applying 1989 characteristics and tuition $2,402 $1,596 72%
Conclusions Tuition increases important, but likely overstated due to data limitations Changes in attainment and demographics explain significant share of debt increase Sizable share of increase unexplained > behavioral change, perhaps shifts in attitudes towards debt?
Next Steps Improve tuition adjustment if possible Repeat analysis using debt to income ratios and other measures of financial distress (e.g., bankruptcy) Do trends in debt spell trouble for the broader student loan market?
Is a Student Loan Cii Crisis on the Horizon? Understanding Changes in the Distribution of Student t Loan Debt over Time Beth Akers, Matthew Chingos, and Alice Henriques Brown Center on Education Policy Brookings Institution Upjohn Institute Student Loans Conference University of Michigan October 26, 2013
Distribution of Education Debt, 1989/1992 and 2010 0 0001 05.0.0000.00015 0 20000 40000 60000 80000 100000 Total Education Debt 1989/1992 2010
Table 2a. Summary Statistics, Household Level, Average Age 20-40 Year Race/Ethnicity of Household Head MaximumEducation of Household Couple White Black Hispanic Other HS or less Some Coll BA Graduate 1989 72% 11% 11% 6% 62% 41% 29% 20% 9% 1992 71% 14% 10% 5% 61% 37% 29% 25% 9% 1995 73% 14% 9% 4% 59% 36% 31% 23% 10% 1998 71% 14% 11% 4% 62% 36% 32% 21% 11% 2001 68% 16% 12% 4% 60% 38% 28% 23% 11% 2004 67% 15% 14% 4% 58% 34% 31% 23% 12% 2007 63% 16% 15% 6% 62% 33% 33% 22% 12% 2010 62% 15% 17% 6% 58% 31% 32% 24% 13%