MAKING SENSE OF LOAN Robert Robert AVERSION AMONG LOW- Kelchen - - PowerPoint PPT Presentation

Sara Sara Goldrick-Rab & Robert

MAKING SENSE OF LOAN

Robert Kelchen U i it f

AVERSION AMONG LOW- INCOME STUDENTS:

University of Wisconsin- Madison

INCOME STUDENTS: EVIDENCE FROM

Michigan

WISCONSIN

2013

$20,000

Trends in ends in Pricing, Public 4- Pricing, Public 4-Year Institutions: 1990-20 ar Institutions: 1990-2013 13

$16,000 $18,000 $12,000 $14,000 $8,000 $10,000 $ , Financial Aid Sticker Price $4 000 $6,000 $8,000 Net Price $0 $2,000 $4,000 $0

OVERVIEW OF LOAN MAXIMUMS

 Stafford Loan limits:

OVERVIEW OF LOAN MAXIMUMS

 Stafford Loan limits:

• $5,500 in year 1 ($3,500 subsidized)
• $6,500 in year 2 ($4,500 subsidized)

$6,500 in year 2 ($4,500 subsidized)

• $7,500 in year 3 and beyond ($5,500 subsidized)

 Total cap=$31,000 for dependent students

• $23,000 subsidized

TRENDS IN BORROWING

 2/3rds of undergrads borrow

TRENDS IN BORROWING

 2/3rds of undergrads borrow

• Average debt per borrower ~$26,000

 Among students at public 4-year institutions, t d t f f ili $30K th t lik l t students from families < $30K are the most likely to borrow

• Average debt per borrower ~$16,500 (approx. 60% of

household income)

• But this is because these students have the most unmet

need

 Among students with unmet need, students from families <$30K are the least likely to borrow

• 57% of them decline to take loans

57% of them decline to take loans

GRADUATION RATES BY FAMILY INCOME GRADUATION RATES BY FAMILY INCOME

Bailey & Dynarski 2011

DEFINING LOAN AVERSION

 Resistance or hesitation to take a loan which would

DEFINING LOAN AVERSION

 Resistance or hesitation to take a loan which would be beneficial to student

• “Beneficial” is especially difficult to define

 Measured by survey questions

• Sometimes with a hypothetical scenario
• Other times, student simply reports “no loans”

 Measured by administrative data

• But nearly always this is “student has no loans”
• But nearly always this is student has no loans

rather than “student declined a loan offered”

• Loan offers are hard to track

PRIOR THEORY AND RESEARCH

 Suggests students with these characteristics may be more

PRIOR THEORY AND RESEARCH

 Suggests students with these characteristics may be more loan averse:

• Racial/ethnic minorities, particularly Hispanic and Asians
• More financial strength

More financial strength

• Less information about the returns to college
• Stronger sense of self-control
• Lower educational expectations
• Lower educational expectations
• Less social capital
• Shorter time horizon
• Inclined to work more
• Inclined to work more

 Loan aversion may also vary by institutional culture  Consequences of loan aversion may vary (see Dwyer)

RESEARCH QUESTIONS

 How does approach to measurement affect who is defined as

RESEARCH QUESTIONS

 How does approach to measurement affect who is defined as loan averse?  What are key demographic disparities in loan aversion?  Which factors drive those disparities?  Which factors drive those disparities?  How is the assessment of the driving factors affected by measurement? H i l i l d d ?  How is loan aversion related to postsecondary outcomes?

METHODS

 Wisconsin Scholars Longitudinal Study

METHODS

 Wisconsin Scholars Longitudinal Study— www.finaidstudy.org  3,000 first-time, full-time Pell recipients attending Wisconsin’s 26 public colleges and universities beginning fall 2008  Sample is 58% female 74% white 80% first  Sample is 58% female, 74% white, 80% first- generation  Average family income ~$29,000

MEASURING LOAN AVERSION

 Utilizing survey and administrative measures from same

MEASURING LOAN AVERSION

Utilizing survey and administrative measures from same time frame: beginning of first year  Survey: “Suppose you could take out a loan of up to $10 000 with a 7 percent interest rate [close to market $10,000 with a 7 percent interest rate [close to market rate]. How much money would you take?”

• Loan averse = student selected $0

 Admin data: We observe all types and amounts of  Admin data: We observe all types and amounts of financial aid offered and accepted

• Loan averse= student declined all loans offered

18% d li d b idi d l

• 18% declined subsidized loans
• 40% declined unsubsidized loans
• (Also, sensitivity checks with different definitions)

MEASUREMENT DIFFERENCES IN LOAN AVERSION

Adminis ministrativ rative Data Data Sur Survey Data Data Borrowed Loan Averse Sur Survey Data Data Borrowed Loan Averse Borrowed 52% 7% Borrowed 52% 7% Loan Averse 29% 12%

N=684; correlation=0.213

INTERPRETATION

 Is the response to a survey question indicative of preference

INTERPRETATION

 Is the response to a survey question indicative of preference

• r intent whereas the administrative data reflect behavior?

 Does this imply that the decision to borrow may sometimes be a constrained choice? a constrained choice?  Or, are these disparities due to timing of when the survey was administered (Sept December 2008) and the timing of when administered (Sept-December 2008) and the timing of when the administrative data was collected (September-October 2008)?  Either way, this implies we ought to have some concern re: measurement

DISPARITIES IN LOAN AVERSION: RACE/ETHNICITY

 Racial/ethnic minority students are generally more loan averse than non-Hispanic white students students

• 11 % pts in admin data but no difference in survey

data; p<.05)  Southeast Asians are more loan averse than non- Hispanic whites 13 % 6 %

• 13 % pts admin data, 6 % pts survey data, p<.05

 Blacks are less loan averse than non-Hispanic whites whites

• ~4 % pts, no measurement difference

DISPARITIES IN LOAN AVERSION:

 No statistically significant differences in loan

OTHER OBSERVATIONS

 No statistically significant differences in loan aversion by gender

• Trends may vary by measurement – survey data says

women are less loan averse while admin data says they are more loan averse  First generation immigrants and students for whom  First-generation immigrants and students for whom English is not their first language are much more loan averse

• 8-16 % pts, more pronounced with admin data

CORRELATES OF LOAN AVERSION: OTHER OBSERVATIONS

Students feel more obligated to support their family and receive no $ support from family while i ll g in college They are less likely to feel college is difficult, and yet also have less high school preparation y g p p Longer time horizons Stronger sense that debt is normative More likely to have received FAFSA help from a non-family member More likely to be working More likely to be working

UNRELATED TO LOAN AVERSION

ACT scores

UNRELATED TO LOAN AVERSION

ACT scores Perceived likelihood of college completion Expected returns to degree (~$60 000) Expected returns to degree (~$60,000) Financial knowledge

• Generalized and specific to financial aid

Generalized and specific to financial aid Locus of control

ACCOUNTING FOR RACIAL/ETHNIC

 Black/white gap measured with survey data

DIFFERENCES IN LOAN AVERSION

 Black/white gap measured with survey data

0.8 1 0.4 0.6

• 0.2

0.2 Student demographics Financial strength Academic preparation Attitudes, beliefs, and Work behaviors Social capital

• 0.6
• 0.4

g p g p p and financial knowledge , dispositions

• 1
• 0.8

ACCOUNTING FOR RACIAL/ETHNIC

 Black/white gap measured with admin data

DIFFERENCES IN LOAN AVERSION

 Black/white gap measured with admin data

0.8 1 0.4 0.6

• 0.2

0.2 Student demographics Financial strength Academic preparation Attitudes, beliefs, and Work behaviors Social capital

• 0.6
• 0.4

de

• g ap cs

st e gt p epa at o and financial knowledge be e s, a d dispositions be a o s

• 1
• 0.8

ACCOUNTING FOR RACIAL/ETHNIC

 Southeast Asian/white gap measured with survey data

DIFFERENCES IN LOAN AVERSION

 Southeast Asian/white gap measured with survey data

0.8 1 0 2 0.4 0.6

• 0.2

0.2 Student demographics Financial strength Academic preparation Attitudes, beliefs, and Work behaviors Social capital

• 0.6
• 0.4

and financial knowledge dispositions

• 1
• 0.8

ACCOUNTING FOR RACIAL/ETHNIC

 Southeast Asian/white gap measured with admin data

DIFFERENCES IN LOAN AVERSION

 Southeast Asian/white gap measured with admin data

0.8 1 0 2 0.4 0.6

• 0.2

0.2 Student demographics Financial strength Academic preparation Attitudes, beliefs, and Work behaviors Social capital

• 0.6
• 0.4

and financial knowledge dispositions

• 1
• 0.8

INSTITUTIONAL CONTEXT AND LOAN AVERSION

Most selectiv Most selective e Some Somewhat what Leas Least selecti t selective 2-ye year colleges colleges Most selectiv Most selective e univ univer ersi sities es Some Somewhat what selectiv selective Leas Least selecti t selective 2-ye year colleges colleges Loan aversion rate 8% 11.1% 24.5% 38.8% % Pell 12-17% 20-38% 23-32% 24% Grad rate 69-81% 41-61% 27-43% 20% % Mi it 6 8 12 8% 5 9 5% 17 22 8% 9 1% % Minority 6.8-12.8% 5-9.5% 17-22.8% 9.1%

ACADEMIC OUTCOMES ASSOCIATED WITH BORROWING

% Enr % Enrolled lled Ye Year 2 % Enr % Enrolled lled Ye Year 3 3 # Credits # Credits Cum.

• um. GPA

Means of Loan Averse Students (Survey data) 78.1 66.9 63.7 2.42 ( y ) Borrowing Impact (survey) 8.9*** 9.3** 4.6** 0.23*** Means of Loan Averse Students (Admin data) 82.4 70.9 64.8 2.58 Borrowing Impact (admin)

• 4.8
• 3.2
• 1.8
• 0.22***

SUMMARY

 Rates of loan aversion vary depending on method of

SUMMARY

 Rates of loan aversion vary depending on method of measurement  Disparities also vary by measurement  Immigration status and home language appear to increase loan aversion A l l i d fi i l  At least among low-income students, financial knowledge, time horizon, etc matter little  Explanations for loan aversion and associations with Explanations for loan aversion and associations with academic outcomes also vary by measurement

IMPLICATIONS

 Is loan aversion a concern? Unclear

IMPLICATIONS

 Is loan aversion a concern? Unclear  Perhaps, given payoff to college  On the other hand, perhaps payoff relative to future earnings is not the relevant comparison With current and future is not the relevant comparison. With current and future earnings committed to family, perhaps debt to household income is more appropriate.

• 24% for families earning < $30K/ year
• 24% for families earning < $30K/ year
• 7% for families earning > $60K/year

 Need to better understand how decisions about borrowing are  Need to better understand how decisions about borrowing are affected by perception that college increases mobility vs. perpetuating economic struggles

• See my upcoming WT Grant study
• See my upcoming WT Grant study