Return on Investment Report

SCHEV's Return on Investment (ROI) Model relates four estimates to compare the cost of a given degree program with its expected return. First, the model compares an estimate of lifetime earnings for holders of that college degree to an estimate of the average expected earnings for holders of a high school diploma. Then the model scales that result with the cost of that degree, which is defined as the sum of student loans and net price. How those values are calculated appears below.

The model starts with constructing a lifetime wage projection model using person-level wage data spanning 25 years, via these five steps:

1. Adjust for inflation: Wage data of graduates from 1995-96 to 2019-20 are adjusted to 2023 values, by each year post-completion for which wages are available, for at least three quarters of the calendar year.
2. Smooth for year-to-year noise: To minimize the impact of random and short-term fluctuations, seven-year moving averages are calculated using median wages per program level and "program area" of instructional program (i.e., two-digit CIP code).
3. Model starting wage and growth: A linear regression model is constructed for each program area at each program level based on historic performance of wages from 1995-96 to 2019-20. For each program level, we limit records to only graduates whose highest degree earned is at the same level.
4. Constrain population: Graduates who have earned both Bachelor’s and Master’s degrees are not included in the construction of the model for Bachelor’s level wage projection.
5. Extrapolate: The model assumes wage changes follow the same pattern beyond 2019-20 in each program area and at each program level and then project a lifetime wage income for graduates individually, using their fourth-year postgraduation wages as an anchor.

Estimating lifetime earnings: Using the 2018-19 Bachelor’s graduating cohort as the study population: Individuals with valid wage data in 2023, and no higher level of degree, are included in the analysis. Assuming these individuals continue to work until age 67 when they reach full retirement under Social Security, the regression model is applied on a by-graduate basis, and their lifetime wage income from 2020 is calculated, until the year they turn 67. In other words, individuals have varying numbers of working years based on their age at graduation, which in turn is based on age at entry and the time spent completing their degree. The longer a student stays in school, typically the lower their lifetime earnings.

Comparison to no college: In parallel, the model constructs a counterfactual lifetime wage projection for the same population where the individuals never entered college, via these three steps:

1. Adjust for inflation: Census Bureau’s ACS 5-year estimates of wages for high school (HS) graduates in Virginia from 2010 to 2023 are adjusted to 2023 values.
2. Estimate start and growth: Linear regression coefficients are calculated based on the above ACS estimates. HS level wages beyond 2023 are projected based on the historic pattern of change in the past 14 years.
3. Extrapolate: Assuming each individual included in the analysis would enter the workforce upon HS graduation, earning HS level wage each year up to age 67, we calculate their HS lifetime wage income. This eliminates the opportunity cost of time in college.

Cost of college: The total cost of a Bachelor’s degree is calculated for each student as the sum of total student loans and net price charged to each student (i.e., institutional budget minus all gift aid in each semester enrolled in a Bachelor’s program). Net prices are adjusted to 2023 equivalents.

• The projected Bachelor’s lifetime wage represents the total gain from earning a Bachelor’s degree.
• The sum of projected high school graduate-level lifetime wages represents the counterfactual comparison.
• The total cost of earning a Bachelor’s degree and years delayed entering the workforce full-time combined represent the real and opportunity costs of earning a Bachelor’s degree.

Contrast gain with loss: The model then contrasts the total gain and total loss. The ROI estimates are represented in two different calculations:

• Net Lifetime Difference=[Projected Bachelor’s Lifetime Wages]-[Projected HS Lifetime Wages]-[ Student Debt]-[Net Price]
  • This is the estimated difference in dollars between college and no college (for this program), subtracting the cost of college.
• ROI ratio=[Projected Bachelor’s Lifetime Wages]/([Projected HS Lifetime Wage]+[Student Debt]+[Net Price])
  • This is the estimated lifetime wages, but divided by or scaled by the comparison of projected HS lifetime wages and college costs.

SCHEV's Return on Investment Model is subject to four limitations:

1. Exclusions: Data are limited to individuals employed in Virginia and in jobs covered by the Commonwealth's Unemployment Insurance program. Certain populations are excluded:
   • Independent Contractors and Gig Workers: Individuals classified as independent contractors rather than employees.
   • Agricultural Laborers: Specific types of farm labor, except where specifically covered.
   • Domestic Service Workers: Individuals working in private homes, local college clubs, or local chapters of fraternities/sororities, unless meeting specific criteria.
   • Religious and Non-Profit Employees: Employees of certain non-profit organizations or religious organizations.
   • Real Estate/Insurance Agents: Agents working solely on commission.
   • Government Officials: Elected officials.
   • Federal Employees: Generally, federal workers are covered under a separate federal system (UCFE), not state UI, including military personnel.
   • Railroad Employees: Covered under the Railroad Unemployment Insurance Act.


2. Underrepresentation: In many programs, graduate or professional study beyond the bachelor’s degree is the primary goal. If students from a certain bachelor’s degree program tend to seek postgraduate degrees in large numbers, the ROI estimates from that degree program will exclude them. This focus on bachelor’s-only may misrepresent the value to those considering the value of that degree if they are considering adding on to it with future postgraduate education. The four-year bachelor’s degree serves a variety of purposes. A degree at this level is designed to demonstrate that the graduate has a broad and general education, combined with a focus on a subject of interest. In the liberal arts and sciences, the area of focus serves to engage the student's interest while sharpening their skills of analysis, inquiry, and communication (among others) while providing a background for further study.


3. Past performance does not guarantee future results: These estimates are based on labor market and wage trends from 1995-2020. This period was marked by the Great Recession of 2007-2009 as well as technological disruption and an expansion of higher education as the proportion of workers with at least a Bachelor’s degree increased from 27% in 2000 to 40% by 2020. These data do not include either the COVID recession or the era of artificial intelligence, which have affected some fields more than others. Using past data to plan one’s future is better than nothing at all, but with the understanding that this history is likely to be not just wrong, but to have certain biases derived from the historical baseline used.


4. Individual characteristics: An individual's outcome and earnings may be affected as much by characteristics of the individual or where they live and work as much or more than as by the degree earned. People tend to live near where they grew up. For example, nearly six in 10 young adults live within 10 miles of where they grew up, and eight in 10 live within 100 miles (Census). Students in Virginia travel a median of 97 miles for college. Upon graduation, some students may have less choice between staying near family or high paying jobs. Further, wage inequalities persist between groups, and some degrees have overrepresentation of some groups and underrepresentation of others.