From the Battlefi eld to the Schoolyard
The Short- Term Impact of the Post- 911 GI Bill
Andrew Barr
Barr
abstract
The Post- 911 GI Bill brought about an enormous expansion in veteran education benefi ts, roughly doubling the average maximum benefi t level
and generating large variation in the magnitude of benefi t expansion across states. Using this context, I explore how enrollment of older, nontraditional
students responds to educational subsidies. The transition from a conditional cash transfer to a more in- kind benefi t affects the relative prices of different
types of education, pushing veterans to invest in more expensive plausibly higher- quality schooling. Using a difference- in- differences strategy,
combined with state level variation in the degree of benefi t expansion, I fi nd that the higher level of benefi ts increased college enrollment of separated
veterans by between 15 and 20 percent while also shifting the composition of enrollment toward four- year schools.
I. Introduction
The Post- 911 Veterans Educational Assistance Act of 2008 Post- 911 GI Bill brought about the largest expansion in veteran education benefi ts since the
end of World War II, roughly doubling the average maximum benefi t level. The size of the benefi t expansion makes it one of the largest increases in fi nancial aid in decades:
Benefi ts pay for in- state tuition, fees, a monthly housing allowance, and a generous stipend for books and supplies.
1
While the fraction of individuals that join the military
1. In a comprehensive overview of fi nancial aid studies, not a single program provided benefi ts equal to or greater than those provided by the Post- 911 GI Bill, and most provided substantially less Deming and
Dynarski 2010.
Andrew Barr is an ABD graduate student in Economics at the University of Virginia. Seminar participants at the University of Virginia have provided many helpful comments. He is grateful to Sarah Turner, Leora
Friedberg, Bill Johnson, and anonymous referees for very helpful suggestions. This research was supported by a grant from the American Educational Research Association, which receives funds for its “AERA
Grants Program” from the National Science Foundation under Grant DRL- 0941014. Opinions refl ect those of the author and do not necessarily refl ect those of the granting agencies. The data used in this
article can be obtained beginning six months after publication through three years hence from Andrew Barr at barrvirginia.edu.
[Submitted May 2013; accepted April 2014] ISSN 0022- 166X E- ISSN 1548- 8004 © 2015 by the Board of Regents of the University of Wisconsin System
T H E J O U R N A L O F H U M A N R E S O U R C E S • 50 • 3
has dropped substantially over the last 60 years, it remains an important component of the population with over 5 percent and nearly 9 percent of males having served
in the military before turning 30. Furthermore, the likelihood of college enrollment for returning veterans is much higher today than after WWII; although roughly 2.2
million of over 16 million WWII veterans attended college during the 12 year duration of the original GI Bill, more than 600,000 of roughly two million recent veterans were
enrolled in college during the 2010–11 school year alone.
Previous studies of federal fi nancial aid programs fi nd mixed evidence of the effec- tiveness of education subsidies in increasing enrollment and largely focus on students
considering enrollment more than two decades ago Hansen 1983, Kane 1995, Seftor and Turner 2002, Dynarski 2003. This study brings new evidence to understanding the
effects of fi nancial aid and, unlike most previous studies, how fi nancial aid impacts the choices of older, nontraditional students.
2
These individuals, who are frequently from low- income backgrounds and among the fi rst in their families to seriously consider
college, are likely those most at risk to underinvest in human capital due to credit or information constraints Barr 2013. I address two fundamental questions: 1 how fi -
nancial aid impacts the enrollment and persistence of these individuals and 2 whether individuals’ school choices respond to a change in the benefi t structure from a pure
conditional cash transfer to a more in- kind benefi t.
In contrast to other research that addresses the effect of fi nancial aid, I am able to use both over time and geographic variation in the level of fi nancial aid to answer
these questions. Unique among the GI Bills and other federal aid programs, the Post- 911 GI Bill benefi ts account for geographic differences in the cost of living and col-
lege attendance. This feature created variation in the size of the benefi t increase across states: Veterans in some states received almost no change in benefi t levels while others
received tens of thousands of dollars in additional benefi ts per year.
I use a difference- in- differences strategy to estimate the impact of the benefi t change on veteran educational choices. Trends in nonveteran enrollment proxy for the
trend that veteran enrollment would have experienced absent the benefi t expansion. I confi rm, through a number of robustness checks, event study analyses, and falsifi ca-
tion exercises, that veteran enrollment trends tracked nonveteran trends leading up to the policy change. Concern remains, however, that veteran enrollment may respond
differently to the Great Recession. While I fi nd no evidence that veteran enrollment is more responsive to labor market contractions, I also leverage cross- state variation
in the degree of benefi t expansion to examine the effect of changing benefi t levels, mitigating these types of concerns. This geographic variation in covered tuition and
housing allowance levels allows for a triple- difference type strategy.
While I spend a considerable amount of time addressing concerns related to the po- tential endogeneity of veteran status later in the paper, I note here two characteristics
of the bill’s passage that mitigate concerns with selection into the military. First, the Post- 911 GI Bill was passed and implemented in roughly 13 months; therefore, the
specifi cs of the bill were not clear until the spring before the implementation. This left
2. Veterans and nontraditional students are similar across a number of characteristics including age, race, motivations for attending a particular institution, institution type, and major; veterans are somewhat more
likely to attend B.A. granting institutions 42 versus 32 percent and more likely to major in business or engineering. Source: Military Service Members and Veterans: A Profi le of those Enrolled in Undergraduate
and Graduate Education in 2007–2008 Radford 2011.
little time for selection into the military to adjust to higher benefi t levels.
3
See Figure 1. Second, the Post- 911 GI Bill benefi t eligibility criteria, standard contract lengths,
and enlistment delays make it unlikely that individuals who enlisted after the bill was announced could have returned and enrolled prior to the end of my sample period.
I fi nd that the expansion of benefi ts increased overall college enrollment of individ- uals with active- duty military service by just under two percentage points. The college
enrollment of separated veterans increased by four to six percentage points a 15 to 20 percent increase. The effects appear to be somewhat larger for males. The enroll-
ment impacts are larger in states that experienced larger benefi t expansions. Overall, the results suggest an increase of at most one percentage point per 1,000 increase in
aid, somewhat smaller than that found by prior researchers. I argue that substantially higher initial levels of aid explain this disparity.
Furthermore, as predicted by the model, the change in the relative prices of different types of education resulted in more veterans enrolling in, and transferring to, four-
3. Figure 1 illustrates that there was little awareness of changes to veterans’ benefi ts until late 2008 at the earliest, and interest remained low until the bill’s implementation.
Figure 1 Interest in the Post- 911 GI Bill
Notes: The vertical lines represent the passage June 30, 2008 and the implementation August 1, 2009 of the Post- 911 GI Bill. Search intensity for each set of keywords is normalized as a ratio relative to the
value in July 2008 the fi rst month with positive search intensities for all queries. Weekly data on search intensity are smoothed using a median spline. Search intensities collected from Google Insight. See Google
Insight documentation for explicit documentation.
year institutions. Finally, I present suggestive evidence that the benefi t expansion has increased the persistence of veterans.
In the next section, I provide a brief overview of the fi nancial aid literature. In Sec- tion III, I describe the Post- 911 GI Bill and the GI Bill that it replaced. In Section IV,
I develop a model of human capital investment to illustrate the changing incentives faced by veterans as the Post- 911 GI Bill went into effect. Section V introduces the
relevant data while Section VI lays out the estimation strategy used to test the implica- tions of the model. Section VII presents the results and Section VIII concludes.
II. Veterans and Education Subsidies