Will Income Share Agreements (ISAs) only work by cherry-picking “STEM” programs? The reality is quite different.
Update (3/2019): I had a great conversation with Will Quist at the 2019 Life Capital conference and acknowledge his well taken point. That is, in the discussion that follows, we are assuming that ISAs have a cap. The question of what kind of cap structure is optimal in ISAs goes beyond the scope of this article. A big thanks to Will for an insightful point!
Income share agreements (ISAs) have started to become a serious financing option for higher education in the last 3 years. Purdue’s experiment with ISAs for its own students is starting to look like a successful model. Other schools, including University of Utah, have announced launching their own ISA programs.
Yet, from the very beginning, there was always the criticism of ISAs as a model that cherry picks higher paying majors and leaving the rest behind. This notion stems from the fact that investors get a percentage of income as an in ISAs. However, the truth is more complicated, and it turns out that ISAs are actually more efficient at the individual and the macro-level.
What are income share agreements?
Lets start from the beginning — what is an income share agreement? In simple terms, an income share agreement is an obligation, but not a loan, where the investor receives a fraction of income from a student during the payment term. The investment is a hybrid investment, in the sense that investors do get upside returns for a range of income levels. The payments are capped, so that once the student completes the payment term or hits a pre-specified payment cap, their obligation is over. Further, if students’ post-graduation income is below a threshold level of income, their obligations are forgiven.
Will ISAs cherry pick high average salary programs like STEM programs?
What kind of students will be selected by ISA investors? It turns out that an ISA’s value depends not just on the average starting salary, but the entire distribution (i.e., spread of the starting salary around its mean).
This latter aspect is important to understand. Because ISA returns are capped, any dispersion in student outcomes will affect ISA investors adversely. In simpler terms, a program with really low placement rates will not work for ISAs. That is, if a significant proportion of students does not receive higher salary outcomes from attending an educational program, then ISAs for such a program are bound to be significant investment losses. The federal government’s Title IV loan programs, in contrast, will provide you the all the loans that you can take at the same interest rate regardless of whether you go to a high placement rate program or a low placement rate program.
What this means is that, given the right investor, a well-designed ISA program can serve programs with lower average starting salaries, so long as graduates are highly likely to get placed (i.e., lower spread around the average). For example, ISAs can work with programs that train teachers and social workers so long as the program has a high placement rate. The caveat is that the cost of the ISA depends on the return required by investors. Impact investors with lower return requirements, and we are working with a few, can bring in much needed capital for such programs. The key takeaway is that ISA programs are more likely to work for programs with higher placement rates and lower average post-graduation salaries than for programs with lower placement rates and higher average post-gradation salaries.
How might ISAs affect overall trends in educational program selection and cost of higher education?
A corollary of the effect described above is that the broader availability of ISAs will make it harder for students to take on educational programs with low placement rates. As this industry matures, with more lower cost capital coming in, I envision that ISA programs will be set up by programs with predictable and high placement rates.
In addition, ISAs will play an important role in controlling the ever-increasing costs of higher education. The Bennett hypothesis says that the availability of low cost federal loans drives up cost of education. Such financing allows people to take on education programs regardless of their ability to repay. As a result, educational programs have a seemingly unlimited demand, in many cases regardless of their ability to impart skills that provide better career outcomes to students. The increasing demand, in turn, drives up education costs.
This notion is actually turned on its head with ISAs. ISAs value programs with better and predictable outcomes for a given cost. Programs with higher costs and less predictable outcomes, that is lower return on investment (ROI), are not good candidates for ISAs. This, in turn means that ISA capital will be primarily available to higher ROI programs, and to the extent that financing directs demand, direct students toward higher ROI programs.