Hypothesis in Search of Evidence: Housing Allowances for Fully Online Post-9/11 GI Bill Beneficiaries

 A March 2023 Urban Institute paper on Post-9/11 housing allowances is a hypothesis in search of evidence, and much of the evidence presented is either wrong, confusing, or misleading.[1], [2] The paper’s authors marshal data from the College Scorecard and student surveys (U.S. Department of Education—ED) and from the GI Bill Comparison Tool (U.S. Department of Veterans Affairs—VA) to argue that the 2011 law granting housing benefits to a class of GI Bill beneficiaries using the Post-9/11 benefit penalizes a cohort of recipients and schools.


  • The paper falsely alleges that Congress failed to explain its rationale for providing a reduced housing benefit to beneficiaries enrolled in fully online programs. In fact, the Congressional Committee report specifically proffered its rationale, explaining that most distance learning is pursued from home with minimal disruption of beneficiaries’ lives, which supports a partial but not a full housing benefit.
  • The report further points out that a full housing benefit was already available to beneficiaries enrolled in programs that offered at least one in-person class, a very liberal policy that had been adopted by VA.
  • Presenting a confusing array of statistics from ED survey data, the paper cherry-picks the Department’s numerous outcome metrics. When small sample sizes undercut the validity of some survey data findings, the paper then focuses on data that includes non-veterans. Regarding outcomes, it makes “apples-to-oranges” evaluations by comparing outcomes between 2-year and 4-year schools.
  • The report fails to identify the institutional sector of schools for which it presents outcome data, neglects to point out the shortcomings of some of the data, and never presents other outcome data that would provide a fuller picture of institutional performance, including graduation rates, median debt after graduation, or percentage of enrollees earning more than a high school graduate.[3], [4]
  • Finally, the paper estimates the cost of providing the full housing allowance to beneficiaries who take all their classes online without disclosing the methodology behind the calculations or noting that the 10-year costs are $1 billion. Ten-year costs are likely a significant barrier to increasing the housing benefit for fully online beneficiaries because it is likely those costs will have to be offset to permit such an increase in spending.

Background on the Post-9/11 GI Bill Housing Benefit

A housing allowance for beneficiaries who took at least one in-person class was available to Post-9/11 beneficiaries before Congress made the decision in 2010 legislation to provide half the national average housing benefit to recipients who attend fully online college programs, which is roughly $988 per month for the academic year 2023.[5], [6], [7], [8] The benefit is paid only when students are actually attending classes; that is, no housing allowance is distributed during Christmas or spring breaks or during the summer unless recipients are enrolled in classes.

In contrast, recipients who attend in-person education programs receive the full housing benefit based on the costs associated with the location of the institution where they physically participate in a majority of classes. This amount differs considerably per school zip code —from several thousand dollars to less than one thousand dollars each month. For example, effective August 1, 2023, recipients enrolled in a college program in New York City will receive $3,666 (with dependents) and $3,297 (no dependents) each month. In contrast, the housing benefit for recipients attending school in Jefferson City, Missouri, will range from $1,071 (with dependents) to $879 (no dependents) monthly.

The generosity of the housing allowance is evidenced by the fact that it represented more than half of the expenditures on the Post-9/11 GI Bill from Fiscal Years 2009 through 2017, eclipsing payments for tuition and fees.

Congress Explained Its Rationale for a Partial Housing Allowance for Fully Online Beneficiaries

In this paper, the authors inaccurately claim that the rationale that led Congress to establish lower housing benefits for fully online beneficiaries compared to students enrolled in in-person programs “…is not well documented, and it is difficult to assess the basis for the disparate treatment….” In fact, Congress did explain its rationale for the disparate treatment of fully online education. The October 2010 Senate Committee on Veterans Affairs report explained its rationale as follows:[9], [10]

The Committee understands that, for some individuals, pursuit of a program of education solely by distance learning may be the only manner available to them and that payment of some portion of the living allowance is appropriate. However, since one of the basic purposes of the living allowance is to offset the cost of housing away from home and since most distance learning is pursued from home, the full allowance does not appear supported at this time.

In short, Congress found that the flexibility of online learning allows beneficiaries to pursue a degree from home without disrupting their lives by moving or giving up their jobs. Failing to acknowledge the lawmakers’ explanation, the paper proposes a series of straw-man hypotheses that essentially add up to a presumed but unsupported bias on the part of lawmakers against beneficiaries that pursue online programs or the outcomes of such programs.

The same Senate Committee report also pointed out that VA already treated a blended experience—a combination of classroom and even one distance learning course—as entirely classroom training, allowing beneficiaries to receive the full housing allowance. As a result, Post-9/11 GI Bill beneficiaries enrolled in fully online programs were only “technically” ineligible for a housing allowance prior to 2011 because of VA’s liberal definition of a blended learning experience.[11], [12]

The Committee recognizes that the trend in higher education appears to be toward the so-called “blended” learning experience where there are components of both classroom instruction and distance learning. This blended experience is provided for under current policies and procedures since VA considers a combination of classroom and even one distance learning course as entirely classroom training.

The authors also conflate the eligibility requirements for the housing allowance. In some paragraphs, they acknowledge that a beneficiary’s allowance varies based on the institution’s location while in others they state that the allowance is based on the student’s area of residence. In fact, the benefit has always been based on an institution’s location.[13], [14]

 Survey Data Challenges                    

Using ED Survey data, the paper concludes that fully online beneficiaries have fewer resources and higher housing costs than those who receive some or all instruction from in-person programs. However, the papers’ authors present various confusing analyses, which suggests that their goal is to overwhelm the reader with statistics in the hope that something “sticks”—that is, convinces the reader that the housing allowance for fully online beneficiaries is unfair. For example, the paper’s authors:

  • present data that involved small sample sizes, which undermine the findings from their analysis;
  • have to acknowledge that that some statistics might be skewed because they include beneficiaries financially dependent on their parents (children of veterans to whom benefits were transferred); and
  • fail to take into account the lack of summer or other routine breaks for online programs compared to in-person programs when reporting average housing allowance costs.

Institutional Outcome Data Require Caveats as Do School Comparisons

A new College Scorecard, which debuted in 2015, contains multiple data points on student outcomes in six categories. The data are on all students at each school; that is, the data do not allow researchers to focus on the outcomes of Post-9/11 beneficiaries. The authors of the paper use two of the College Scorecard data points—earnings and progress in paying down student loans. They fail to disclose limitations in these data points that paint a more complete picture of student outcomes at 48 institutions “popular among GI Bill recipients.”

Data Discrepancies

The paper’s data analysis of earnings is sloppy and sometimes misleading. Data for several schools appear to be misstated, and fully online enrollment for many schools is inconsistent with that reported by ED on its College Navigator website, which is the source of College Scorecard data. For example:

  • Trident at American InterContinental University is not included in the Scorecard and presumably the authors used data for American InterContinental University, which is included in the Scorecard. Both schools offer essentially fully online programs. When Perdoceo, the company that owns American InterContinental, purchased Trident University International, it maintained Trident as an independent entity.
  • It is unclear if the authors conflated Arizona State data for in-person programs with its fully online option. The authors provide data for the Tempe, Arizona, campus, which is in-person instruction, and claim that fully online enrollment is 34 percent. However, College Navigator shows fully online enrollment for the Tempe campus as 6 percent. Arizona State’s fully online “digital immersion” program (100 percent) is in nearby Scottsdale.
  • The University of Arizona is listed twice by the authors—both its in-person programs (University of Arizona) and its fully online programs (University of Arizona Global Campus). However, University of Arizona Global Campus is not listed in College Scorecard. Rather the information reported is from Ashford University, which the University of Arizona purchased in 2020 and rebranded as University of Arizona Global Campus.
  • Embry-Riddle Online is reported by the authors as having 38 percent enrollment in distance learning, but College Navigator reports it at 90 percent.
  • The authors similarly understate fully online enrollment at Austin Community College Districts and Lone Star College system by 31 and 15 percentage points respectively.
  • The University of Texas at Arlington’s fully online enrollment is reported by the authors as 15 percentage points higher than on College Navigator.

School Sector Not Disclosed

The authors’ analysis fails to disclose the for-profit ties of the fully online programs at the University of Arizona’s and Purdue University. Moreover, several for-profit schools settled or lost lawsuits for misleading advertising and recruiting with federal or state authorities.

  • The University of Arizona purchased Ashford University’s fully online programs from Zovio in 2020, and in 2022 Zovio lost a lawsuit filed by the Attorney General of California in 2017 for misleading advertising and recruiting. Zovio was fined $22 million. Based on similar fraudulent conduct, Zovio is also on the hook to ED for student loan debt cancellation under the Borrower Defense rule. Although the University of Arizona is a public sector institution, ED lists the University of Arizona Global Campus as a for-profit entity under its previous name, Ashford University.[15]
  • Purdue University’s purchase of for-profit Kaplan University was finalized in 2018 and Kaplan University was rebranded as a fully online public sector institution. According to a Purdue press release, Kaplan provides “back-office services” to Purdue Global, such as marketing and advertising, front-end student advising, admissions support, and financial aid and student finance. Kaplan, Inc., the owner of Kaplan University, had settled several lawsuits for misleading advertising and recruiting prior to the sale.

Other for-profit schools which the authors claim are “popular” with GI Bill beneficiaries negotiated multiple settlements for misleading advertising and recruiting include the University of Phoenix (Apollo Education Group), DeVry University, and Career Education Corporation (now known as Zovio and the owner of American InterContinental University, Trident University International, and Colorado Technical University).

Scorecard Data Limitations and Excluded Data

Each data point on College Scorecard defines who is included or excluded in the data presented, which the paper’s authors do not disclose. Moreover, other important outcome data that would add important context to their analysis is excluded.

  • The median earnings data used by the paper’s authors tracks a cohort of students who enrolled at the same time and relies on tax data to report their income 10 years later. Although the data includes both individuals who graduated and those who did not, students who have a deferment (not working or still enrolled in a school) are excluded, raising a cautionary note about the earnings data, which the authors do not disclose.
  • The authors also do not report another important outcome metric—the proportion of students earning more than a high school graduate.
  • While the authors analyze but do not report Scorecard data for the 48 schools showing the proportion of students making progress in paying off federal student loans, they fail to mention other important Scorecard metrics such as average annual cost, loans in forbearance or deferment, loans in default, and median federal debt after graduation. GAO found that schools carefully manage their default rates, encouraging students to defer loan payments in order to avoid defaulting, which raises a question about the accuracy of the default rate data.
  • Graduation and retention data on the Scorecard, which would have added important context to the earnings and debt statistics was also ignored by the paper’s authors, including the proportion of students who withdrew or transferred.

Apples to Oranges Comparisons

The authors fail to identify whether the institutions whose data they analyze are 4-year institutions that offer bachelor’s degrees or higher or 2-year colleges that offer only associate degrees or certificates. As a result, their comparisons of earnings or student loan debt repayment are misleading.

  • Earnings. The authors compare median earnings at Texas A&M University ($66,566), a 4-year public university to those at Tidewater Community College, a 2-year community college ($35,053). This is misleading because earnings for an associate’s degree, such as Tidewater, are lower than for a bachelor’s degree, the typical degree earned at Texas A&M. According to a study by Georgetown University’s Center on Education and the Workforce, bachelor’s degree holders earn 31 percent more than those with an associate’s degree. Moreover, graduation rates at 4-year (83 percent at Texas A&M) schools are generally higher than at community colleges (20 percent) where many students either withdraw or transfer to a 4-year institution.[16], [17]
  • Student loan debt repayment. Compared to Tidewater, a 2-year institution, 4-year colleges such as Texas A&M are more expensive (more than twice as costly), take longer to earn a degree, and graduate students with more debt but with a higher proportion making progress paying off their student loans.

Increase in Housing Allowance for Beneficiaries Fully Online Would Likely Require Budgetary Offsets

Without disclosing its methodology, the paper projects that the cost of providing the same housing allowance to fully online beneficiaries that individuals receive at in-person programs would be $100 million each year. The paper fails to point out that the estimated $100 million annual cost would total $1 billion over 10 years, a significant increase if veteran education budgetary offsets (benefit cuts) are required.[18] For example, the increased cost of providing Post-9/11 GI Bill benefits to the National Guard, reservists, and non-degree seeking students in 2011 was offset by eliminating the housing allowance during routine pauses in college classes, such as for Christmas or spring breaks. It is also important to acknowledge that the proportion of the Post-9/11 GI Bill’s costs attributable to the housing allowance from fiscal years 2009 through 2017 accounted for 54 percent of the program’s costs.[19]

[1] Delisle, Jason and Jason Cohen, Urban Institute, “Housing Allowances and the GI Bill: Should Veterans Enrolled Online Receive Full Benefits?” (March 2023), https://www.urban.org/sites/default/files/2023-03/Housing%20Allowances%20and%20the%20GI%20Bill.pdf.
[2] The report was underwritten by Veterans Education Project, an organization believed to receive funding from for-profit education companies that run large online programs.
[3] A 2018 report by Veterans Education Success describes the difficulty of using the Scorecard data to make comparisons between different institutions and contains a glossary that describes each outcome measure as of 2018.
[4] Veterans Education Success, “Association of Career Education Colleges and Universities: Outcomes at Selected Member Schools” (2018), https://vetsedsuccess.org/career-education-colleges-universities-outcomes-selected-members-schools/. (Note that the Department of Education has made changes to the data since 2018.)
[5] See Section 102(b)(2) of the Post-9/11 Veterans Educational Assistance Improvements Act of 2010, Pub. L. No. 111-377, 124 Stat. 4106 (2011), wherein it adds the following stipulation for fully online distance learning, “(iii) In the case of an individual pursuing a program of education solely through distance learning on more than a half-time basis, a monthly housing stipend equal to 50 percent of the amount payable under clause (ii) if the individual were otherwise entitled to a monthly housing stipend under that clause for pursuit of the program of education.”
[6] Post-9/11 Veterans Educational Assistance Improvements Act of 2010. U.S.C. § 102(b)(2) (2011).
[7] The housing benefit mirrors the basic housing allowance that the U.S. Department of Defense provides to active duty servicemembers, a benefit that is adjusted for inflation annually each January. For Post-9/11 beneficiaries, the rate is effective for the academic year that begins on August 1 of each year and is updated at that time. On August 1, 2023, the housing allowance for fully online programs will increase from $938.50 to $988, which is half the national average of the benefit.
[8] U.S. Department of Veterans Affairs, “Post-9/11 GI Bill (Chapter 33): How does VA determine my monthly housing allowance (MHA)?” (2023), www.va.gov/education/about-gi-bill-benefits/post-9-11/#how-does-va-determine-my-month.
[9] The benefit was included in P.L. 111- 377, § 102 (b)(iii), the Post-9/11 Veterans Educational Assistance Improvement Act of 2010.
[10] Post-9/11 Veterans Educational Assistance Improvements Act of 2010. U.S.C. § 102(b)(2) (2010).
[11] As Veterans Education Success reported in 2018, several public and nonprofit online schools were taking advantage of VA’s extremely liberal definition of classroom training and recruiting beneficiaries by offering online programs with only one in-person class, thereby qualifying enrollees for the full housing benefit. Ironically, the Urban Institute paper suggests that the reduced housing allowance policy for fully online programs “could also encourage students to needlessly enroll in just one in-person class while taking their other classes online simply to increase their MHA [military housing allowance].” VA policy already encouraged such behavior, and it is unclear why taking one online class to receive the full housing allowance is a burden on most beneficiaries.
[12] Veterans Education Success, “Association of Career Education Colleges and Universities: Outcomes at Selected Member Schools” (2018), https://vetsedsuccess.org/wp-content/uploads/2019/01/career-education-colleges-universities.pdf.
[13] Until 2017, payments were based on the zip code of the institution where the student’s enrollment was certified (see § 106 of the Harry W. Colmery Veterans Education Assistance Act of 2017). However, this methodology allowed some schools to game the payments because enrollment was certified in a high-cost zip code even though beneficiaries were taking their in-person classes in a zip code with significantly lower housing allowances. As a result, other institutions in those lower-cost zip codes were at a competitive disadvantage in attracting Post-9/11 beneficiaries. In 2017, Congress changed the basis for determining the housing allowance to the zip code of the campus in which a beneficiary “physically participates in a majority of classes.”
[14] The Harry W. Colmery Veterans Education Assistance Act, Title 38 U.S.C. § 106. (2017), https://www.congress.gov/congressional-report/115th-congress/house-report/247/1.
[15] U.S. Department of Education. Integrated Postsecondary Education Data System. Accessed June 13, 2023. https://nces.ed.gov/ipeds/datacenter/institutionprofile.aspx?unitId=154022.
[16] The midpoint of the graduation rate for all graduates at 2-year and 4-year schools is also quite different, 32 percent and 58 percent respectively.
[17] Carnevale, A., Rose, S., & Cheah, B, Georgetown University, “The College Payoff: Education, Occupations, Lifetime Earnings,” (Aug. 13, 2021), https://cew.georgetown.edu/cew-reports/the-college-payoff/.
[18] Under the 2010 Statutory Pay-As-You Go (PAYGO) Act, any legislative changes to taxes or mandatory spending that increases projected deficits must be “offset” or paid for by other changes to taxes or mandatory spending that reduce deficits by an equivalent amount. (The House and Senate each enforce the PAYGO principle though similar internal rules, independent of the Congressional Budget Act, although those rules can be waived by majority vote in the House and by the vote of 60 senators.) For more details, see Center on Budget and Policy Priorities, “Policy Basics: Introduction to the Federal Budget Process” (2022, October), https://www.cbpp.org/research/federal-budget/introduction-to-the-federal-budget-process.
[19] See Veterans Education Success, “GI Bill Payments and Enrollment, Fiscal Years 2009 to 2018,” (Oct. 2019), /https://vetsedsuccess.org/gi-bill-payments-and-enrollment-fiscal-years-2009-to-2018/.

MHA Housing Allowance.vFINAL