Aid 
Policy: 
Lessons 
from 
Research 
Susan 
Dynarski, 
University 
of 
Michigan 
Judith 
Sco@-­‐Clayton, 
Columbia 
University
Outline 
1. Higher 
EducaIon 
Structure 
in 
US 
2. SIcker 
Price 
vs. 
Net 
Price 
3. Evidence 
on 
the 
Effect 
of 
Price 
2
Provision 
of 
Postsecondary 
EducaIon 
in 
US 
• Public 
Sector: 
80% 
of 
undergrads 
• Private, 
non-­‐profit: 
10% 
• Private, 
for-­‐profit: 
10%
Provision 
of 
Postsecondary 
EducaIon 
in 
US 
• Public: 
80% 
of 
undergrads 
– UniversiIes, 
colleges 
• Four-­‐year, 
BA 
– Community 
colleges 
• Two-­‐year, 
AA 
• Private, 
non-­‐profit: 
10% 
– 
UniversiIes, 
colleges 
• Private, 
for-­‐profit: 
10%
Funding 
of 
Postsecondary 
InsItuIons 
in 
US 
• Public 
Sector 
– States 
give 
$ 
to 
schools 
– Schools 
charge 
low 
tuiIon 
• Private, 
non-­‐profit 
– 
Some 
have 
large 
endowments 
– Schools 
charge 
high 
tuiIon 
• Private, 
for-­‐profit 
– Schools 
charge 
high 
tuiIon
6 
$30,000 
$25,000 
$20,000 
$15,000 
$10,000 
$5,000 
$0 
InflaIon-­‐Adjusted 
TuiIon 
and 
Fees 
Private 
Nonprofit 
Four-­‐Year 
Public 
Four-­‐Year 
Public 
Two-­‐Year 
(2011$) 
1981-­‐82 
1986-­‐87 
1991-­‐92 
1996-­‐97 
2001-­‐02 
2006-­‐07 
2011-­‐12 
SOURCE: 
The 
College 
Board, 
Trends 
in 
College 
Pricing 
2012.
Public 
Prices 
Rise 
as 
State 
AppropriaIons 
Fall 
SOURCE: 
The 
College 
Board, 
Trends 
in 
College 
Pricing 
2012, 
Figure 
12A.
8 
Average 
State 
AppropriaIons 
for 
Higher 
EducaIon 
per 
$1,000 
in 
Personal 
Income 
SOURCE: 
The 
College 
Board, 
Trends 
in 
College 
Pricing 
2012, 
Figure 
13A.
SIcker 
Price 
vs. 
Net 
Price 
9
Aid 
for 
College 
• Federal 
government 
provides 
over 
$180 
billion 
in 
support 
for 
over 
20 
million 
students 
annually 
– $80 
billion 
is 
non-­‐repayable 
(i.e 
not 
loans) 
– 9 
million 
Pell 
Grants 
($35 
billion 
total) 
– 18 
million 
tax 
credits 
($19 
billion 
total) 
• State 
governments 
provide 
aid 
to 
students 
• Colleges 
provide 
aid 
to 
students
Net 
Price 
≠ 
SIcker 
Price 
SOURCE: 
Based 
on 
data 
in 
The 
College 
Board, 
Trends 
in 
College 
Pricing, 
mulIple 
years. 
14
Sources: 
NPR 
Planet 
Money 
graphic 
based 
on 
data 
from 
College 
Board, 
Trends 
in 
College 
Pricing
EsImaIng 
the 
Effect 
of 
Price 
on 
Student 
Decisions
EsImaIon 
Issues 
• We 
would 
like 
to 
know 
how 
the 
price 
of 
postsecondary 
educaIon 
affects 
schooling 
decisions 
• What 
is 
the 
effect 
of 
a 
$1000 
change 
in… 
– TuiIon 
price? 
– Aid? 
• Relevant 
for 
both 
policy 
and 
research
How 
does 
the 
aid 
affect 
schooling 
decisions?
Why 
does 
aid 
eligibility 
vary? 
• Need-­‐based 
aid 
• Merit 
aid 
• AthleIc 
scholarships
Causal 
InterpretaIon? 
• Unobserved 
variables 
are 
correlated 
with 
aid 
and 
schooling 
outcomes 
à 
omi@ed-­‐variables 
bias 
• SoluIon: 
randomized 
trial 
or 
natural 
experiment
Quasi-­‐Experimental 
Analysis: 
Social 
Security 
Student 
Benefits 
• Paid 
sIpends 
to 
college-­‐going 
children 
of 
deceased, 
disabled 
and 
reIred 
Social 
Security 
beneficiaries 
– Established 
1965 
• Scope 
of 
Program 
(1981) 
– $3.3 
billion 
– 850,000 
students 
– 1 
in 
10 
college 
students 
– 2nd 
largest 
federal 
student 
grant 
program 
• Generosity 
of 
Benefits 
– averaged 
$3000/year 
(nominal) 
• average 
public 
university 
tuiIon: 
$1000 
• average 
private 
university 
tuiIon: 
$5000
Social 
Security 
Student 
Benefit 
Program 
• Program 
Eliminated 
1981 
• Students 
not 
in 
college 
as 
of 
May, 
1982 
ineligible 
for 
benefits 
• Spending 
dropped 
from 
$3.3 
billion 
to 
$0.3 
billion 
in 
two 
years 
• One 
of 
the 
largest 
policy 
innovaIons, 
ever, 
in 
student 
aid
Father Not 
Deceased 
=control 
group 
Father 
Deceased 
=eligible 
group 
Difference 
High School Senior 1979-81 
=before 
YC0 YT0 YT0 - YC0 
High School Senior 1982-83 
=after 
YC1 YT1 YT1 - YC1 
Difference-in-Difference (YT1 - YC1) – 
(YT0 - YC0) 
Difference-­‐in-­‐Differences
InterpreIng 
Size 
of 
Effects 
• Eligibility 
for 
program 
increased 
– college 
entry 
22 
percentage 
points 
– completed 
schooling 
0.68 
years 
• Offer 
of 
$1000 
in 
student 
benefits 
increases 
– college 
entry 
by 
3.6 
percentage 
points 
– completed 
schooling 
by 
0.16 
years
Lessons 
from 
Evidence 
on 
the 
Effect 
of 
Price 
on 
Student 
Decisions
30 
Lesson 
1: 
Money 
ma@ers 
for 
college 
entry 
• Several quasi-experimental studies find that 
when prices decreases, enrollment increases 
• Roughly 3-5 percentage point increase in 
enrollment per $1,000 in grant aid 
• Evidence of effects of grant aid on completion is 
less conclusive
31 
Lesson 
2: 
Complexity 
undermines 
effecIveness 
• Funding system in US is extremely complicated 
– Huge variation in tuition prices 
– Dozens of aid programs 
• Widespread misinformation about net price of college 
– Overestimates of tuition prices 
– Underestimates of aid 
• Randomized trial: complete federal aid application for 
potential students 
– Large increases in college attendance, persistence 
• Efforts underway to simplify aid programs
Aid 
ApplicaIon
36 
Lesson 
3: 
IncenIves 
increase 
effecIveness 
of 
aid 
• Aid can be unconditional or conditional on performance 
• Evidence on conditional aid programs indicates they 
increase college persistence and completion 
• The conditions may signal appropriate levels of effort 
– Number of courses to take 
– Acceptable grades
37 
Lesson 
4: 
Evidence 
on 
loans 
is 
limited 
• Shockingly limited evidence on the effects of 
loans given their prominence in college finance 
• Students clearly don’t like them and they don’t 
appear to be as effective as grants 
• But are they cost-effective compared to grants? 
• So much attention has been focused on interest 
rates, but other features of loan repayment may 
be even more important
Other 
findings 
• Little evidence that financial aid drives up tuition 
prices (the “Bennett Hypothesis”), except for in 
the for-profit sector 
• Some evidence that Pell grants “crowd out” 
institutional aid at selective private institutions, 
but not at the public institutions enrolling majority 
of recipients 
38

Aid Policy: Lessons from Research

  • 1.
    Aid Policy: Lessons from Research Susan Dynarski, University of Michigan Judith Sco@-­‐Clayton, Columbia University
  • 2.
    Outline 1. Higher EducaIon Structure in US 2. SIcker Price vs. Net Price 3. Evidence on the Effect of Price 2
  • 3.
    Provision of Postsecondary EducaIon in US • Public Sector: 80% of undergrads • Private, non-­‐profit: 10% • Private, for-­‐profit: 10%
  • 4.
    Provision of Postsecondary EducaIon in US • Public: 80% of undergrads – UniversiIes, colleges • Four-­‐year, BA – Community colleges • Two-­‐year, AA • Private, non-­‐profit: 10% – UniversiIes, colleges • Private, for-­‐profit: 10%
  • 5.
    Funding of Postsecondary InsItuIons in US • Public Sector – States give $ to schools – Schools charge low tuiIon • Private, non-­‐profit – Some have large endowments – Schools charge high tuiIon • Private, for-­‐profit – Schools charge high tuiIon
  • 6.
    6 $30,000 $25,000 $20,000 $15,000 $10,000 $5,000 $0 InflaIon-­‐Adjusted TuiIon and Fees Private Nonprofit Four-­‐Year Public Four-­‐Year Public Two-­‐Year (2011$) 1981-­‐82 1986-­‐87 1991-­‐92 1996-­‐97 2001-­‐02 2006-­‐07 2011-­‐12 SOURCE: The College Board, Trends in College Pricing 2012.
  • 7.
    Public Prices Rise as State AppropriaIons Fall SOURCE: The College Board, Trends in College Pricing 2012, Figure 12A.
  • 8.
    8 Average State AppropriaIons for Higher EducaIon per $1,000 in Personal Income SOURCE: The College Board, Trends in College Pricing 2012, Figure 13A.
  • 9.
    SIcker Price vs. Net Price 9
  • 12.
    Aid for College • Federal government provides over $180 billion in support for over 20 million students annually – $80 billion is non-­‐repayable (i.e not loans) – 9 million Pell Grants ($35 billion total) – 18 million tax credits ($19 billion total) • State governments provide aid to students • Colleges provide aid to students
  • 14.
    Net Price ≠ SIcker Price SOURCE: Based on data in The College Board, Trends in College Pricing, mulIple years. 14
  • 15.
    Sources: NPR Planet Money graphic based on data from College Board, Trends in College Pricing
  • 16.
    EsImaIng the Effect of Price on Student Decisions
  • 17.
    EsImaIon Issues •We would like to know how the price of postsecondary educaIon affects schooling decisions • What is the effect of a $1000 change in… – TuiIon price? – Aid? • Relevant for both policy and research
  • 18.
    How does the aid affect schooling decisions?
  • 19.
    Why does aid eligibility vary? • Need-­‐based aid • Merit aid • AthleIc scholarships
  • 20.
    Causal InterpretaIon? •Unobserved variables are correlated with aid and schooling outcomes à omi@ed-­‐variables bias • SoluIon: randomized trial or natural experiment
  • 21.
    Quasi-­‐Experimental Analysis: Social Security Student Benefits • Paid sIpends to college-­‐going children of deceased, disabled and reIred Social Security beneficiaries – Established 1965 • Scope of Program (1981) – $3.3 billion – 850,000 students – 1 in 10 college students – 2nd largest federal student grant program • Generosity of Benefits – averaged $3000/year (nominal) • average public university tuiIon: $1000 • average private university tuiIon: $5000
  • 22.
    Social Security Student Benefit Program • Program Eliminated 1981 • Students not in college as of May, 1982 ineligible for benefits • Spending dropped from $3.3 billion to $0.3 billion in two years • One of the largest policy innovaIons, ever, in student aid
  • 24.
    Father Not Deceased =control group Father Deceased =eligible group Difference High School Senior 1979-81 =before YC0 YT0 YT0 - YC0 High School Senior 1982-83 =after YC1 YT1 YT1 - YC1 Difference-in-Difference (YT1 - YC1) – (YT0 - YC0) Difference-­‐in-­‐Differences
  • 28.
    InterpreIng Size of Effects • Eligibility for program increased – college entry 22 percentage points – completed schooling 0.68 years • Offer of $1000 in student benefits increases – college entry by 3.6 percentage points – completed schooling by 0.16 years
  • 29.
    Lessons from Evidence on the Effect of Price on Student Decisions
  • 30.
    30 Lesson 1: Money ma@ers for college entry • Several quasi-experimental studies find that when prices decreases, enrollment increases • Roughly 3-5 percentage point increase in enrollment per $1,000 in grant aid • Evidence of effects of grant aid on completion is less conclusive
  • 31.
    31 Lesson 2: Complexity undermines effecIveness • Funding system in US is extremely complicated – Huge variation in tuition prices – Dozens of aid programs • Widespread misinformation about net price of college – Overestimates of tuition prices – Underestimates of aid • Randomized trial: complete federal aid application for potential students – Large increases in college attendance, persistence • Efforts underway to simplify aid programs
  • 32.
  • 36.
    36 Lesson 3: IncenIves increase effecIveness of aid • Aid can be unconditional or conditional on performance • Evidence on conditional aid programs indicates they increase college persistence and completion • The conditions may signal appropriate levels of effort – Number of courses to take – Acceptable grades
  • 37.
    37 Lesson 4: Evidence on loans is limited • Shockingly limited evidence on the effects of loans given their prominence in college finance • Students clearly don’t like them and they don’t appear to be as effective as grants • But are they cost-effective compared to grants? • So much attention has been focused on interest rates, but other features of loan repayment may be even more important
  • 38.
    Other findings •Little evidence that financial aid drives up tuition prices (the “Bennett Hypothesis”), except for in the for-profit sector • Some evidence that Pell grants “crowd out” institutional aid at selective private institutions, but not at the public institutions enrolling majority of recipients 38