Two decades of debt: How has the profile of student loan borrowers changed from 1992 to 2012? Nicholas Hillman University of Wisconsin-Madison AIR 2016 @n_hillman This material is based upon work supported by the Association for Institutional Research, the National Science Foundation, and the National Center for Education Statistics under Association for Institutional Research grant number RG15-9327. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views 1 of the Association for Institutional Research, the National Science Foundation, or the National Center for Education Statistics
Context Why is debt rising? Enrollment growth Shrinking family income State divestment Policy choices How do we measure debt? Borrowing vs repaying Annual vs cumulative Graduates vs non-completers NSLDS vs self-reports What do we know? Impacts of debt are mixed Enrollment patterns matter Distribution of debt is unclear What don t we know? Who borrows? How much? How has this changed over time? @n_hillman Context Design Findings Next Steps 2
Design Sample and measures National Postsecondary Student Aid Study: 1992-2012 All undergraduates public 2yr, public 4yr, non-profit 4yr, for-profit Outcome: annual Federal Title IV loans excluding PLUS Controls: sector, tuition, class level, gender, age, dependency, race, income Double hurdle regression Participation stage: did you borrow? Consumption stage: how much did you borrow? Pooled with year interactions @n_hillman Context Design Findings Next Steps 3
Less surprising Upperclassmen are more likely to borrow and borrow more Participation and consumption levels vary by sector Later waves are more likely to borrow and borrow more More surprising Participation and consumption dip during 2000s Consumption is harder to predict than participation Black students (not students of color ) are more likely to borrow and borrow more Shift towards upper-income borrowing 4 @n_hillman Context Design Findings Next Steps 4
1992 1996 2000 2004 2008 2012 1992 1996 2000 2004 2008 2012 Findings 100% Public 2yr Public 4yr $8,000 75% $6,000 50% $4,000 25% $2,000 0% $0 100% $8,000 75% $6,000 50% $4,000 25% $2,000 0% Non-profit 4yr For-profit $0 @n_hillman Context Design Findings Next Steps 5
1992 1996 2000 2004 2008 2012 1992 1996 2000 2004 2008 2012 Findings 100% Public 2yr Public 4yr $8,000 75% $6,000 50% $4,000 25% $2,000 0% $0 100% $8,000 75% $6,000 50% $4,000 25% $2,000 0% Non-profit 4yr For-profit $0 @n_hillman Context Design Findings Next Steps 6
Q1: Who borrows? In 1992, race/ethnicity was not a predictor of participation By 2000, this all changed and race became statistically significant 0.50 0.40 0.30 0.20 Black White Ameri. Indian Hispanic Asian/PI 0.10 Gaps in participation are widening over time 0.00 1992 1996 2000 2004 2008 2012 @n_hillman Context Design Findings Next Steps 7
0.50 0.40 0.30 0.20 0.10 0.00 1992 1996 2000 2004 2008 2012 Q1: Who borrows? Lowest two income groups no longer most likely to borrow Top three income groups increased faster than bottom two Shift towards upper-income borrowers over time @n_hillman Context Design Findings Next Steps 8
0.50 0.40 0.30 0.20 0.10 0.00 1992 1996 2000 2004 2008 2012 Q1: Who borrows? Lowest two income groups no longer most likely to borrow Top three income groups increased faster than bottom two Shift towards upper-income borrowers over time @n_hillman Context Design Findings Next Steps 9
Q2: How much do students borrow annually? Average debt over time Doubled if we don t adjust Rose by 25% increase if we do Or about 1.25% annual growth $7,000 $6,000 $5,000 $4,000 $3,000 $2,000 $1,000 $0 1992 1996 2000 2004 2008 2012 non-adjusted adjusted @n_hillman Context Design Findings Next Steps 10
Q2: How much do students borrow? Black students borrowed less than white students prior to 2000. $7,000 $6,500 $6,000 $5,500 $5,000 Other racial/ethnic differences not $4,500 statistically significant $4,000 1992 1996 2000 2004 2008 2012 @n_hillman Context Design Findings Next Steps 11
Q2: How much do students borrow? $7,000 $6,500 Income differences are not statistically different Until 2012, when highest income students began to borrow significantly more than lower income students $6,000 $5,500 $5,000 $4,500 $4,000 1992 1996 2000 2004 2008 2012 @n_hillman Context Design Findings Next Steps 12
Less surprising Upperclassmen are more likely to borrow and borrow more Participation and consumption levels vary by sector Later waves are more likely to borrow and borrow more More surprising Participation and consumption dip during 2000s Consumption is harder to predict than participation Black students (not students of color ) are more likely to borrow and borrow more Shift towards upper-income borrowing 13 @n_hillman Context Design Findings Next Steps 13
Next Steps Aiming for parsimony, but maybe impossible here Building toward a typology of borrowers Primarily focusing on analysis at this point in time Pooled vs. just using 2012 and 1992 (or 2016 and 1996) Cumulative debt versus annual debt Focus on subsamples By sector By level By dependency By loan type @n_hillman Context Design Findings Next Steps 14