Government Income Share Agreements: A Unique Proposal For Fixing Public Education

10 minute read

I recently stumbled on an interesting and novel proposal for how governments can better fund mass education, while making access more equitable, and promoting better education outcomes.

The idea comes from a book called Investing In Human Capital (published in 2004) by economist Miguel Palacios. The book is notable for helping to develop alternative models of student finance, such as income contingent loans and income share agreements. These models are now increasingly being applied by universities, coding bootcamps, and by alternative student finance companies. A common theme of these models is that the amount paid back by the student is dependent on how much the student earns after graduating.

However, one of the ideas proposed by Palacios is a bit different, in that the student does not have to pay back anything at all. Instead, the financier is paid back by government directly from the tax revenue collected from that student down the line. This particular model covers only a few pages of the book, but it’s an idea I keep coming back to. It is elegant, maybe a bit radical, and it seems to have completely fallen through the cracks of history. I cannot find any discussion or reference to this model outside of the book (if anyone has seen this idea elsewhere please correct me on this point). I feel it deserves to see the light of day, hence this post.

Government Income Share Agreements

The basic idea:
Government invites private investors to finance the education of tertiary learners.1 In exchange, government commits to repaying the investor a pre-agreed share of the future tax revenue which it (government) ends up collecting from that learner (if and when the learner enters the working economy). Every student funded gives rise to a Government Income Share Agreement (or “GISA”)2 between the financier and government.

Let’s explore some of the details:

Any student that receives funding from this initiative is not out-of-pocket in any way and has to pay nothing back to the financier. The repayment to the financier comes out of the income tax, which the student will pay anyway when they enter the working economy.

The funders do not choose specific individuals to fund. Instead they fund all students in a specific course or college (I use the word “college” to refer to any type of institution that provides tertiary education). Funders are essentially betting on certain courses to produce successful learners. If their bet is successful, the government wins, because the government has a new batch of taxpayers. As such, the government commits to sharing some of that upside with the investor in order to reward the bet.

The amount owed by government is a specified percentage of the taxable income of the student. If the country has a well-functioning tax system, then the investor can be confident that the earnings of the student will be tracked and repaid. The more the students ends up earning the higher the repayments (however the total repayments can be capped at a certain amount).

As a result, the government gets to keep less income tax from the learner. But, perhaps, had the learner not benefited from the funding, he or she would be paying little or no income tax at all.

In theory it is a win-win all round. The investor gets a return on investment. The student gets a better education. And the government gets a more educated and productive citizen (which of course also benefits society at large).

The benefit of this model is twofold; Firstly, it provides new sources of funding. Secondly, the particular repayment criteria is such that it helps to ensure that funds are spent optimally.

How do GISA’s deliver education to those who most need it?

Investors would therefore want to fund the best and brightest students with the best career prospects. To some extent this is fine. Normal government models for tertiary funding also take into account how well a student did in school before the student is funded. But these government models have various mechanisms to ensure a measure of equality. Such mechanisms would also need to be enforced in a GISA system as private investors might otherwise prioritise students from privileged backgrounds who are deemed more likely to succeed and thus less risky an investment.

Fortunately, government has little incentive in taking on funding for privileged learners who already have good career prospects - why give away future tax revenue when that revenue is already pretty much guaranteed?

Instead, the incentive for Government is in taking on investment for poor learners who would otherwise struggle to get a good education. If those students can be well educated and prepared for the workforce, then this is bonus tax revenue for the government, and it is worth parting with some of that tax revenue if someone is willing to take on the cost and risk of funding the student.

Attracting finance

If government only seeks funding for the students with the least prospects, would investors be interested at all? How do Investors get a return on students with poor career prospects?

Firstly, government may have to price the risk into the deal and offer premium returns to investors. However, this does not diminish the attractiveness of the deal for government. Having uneducated adults is a drain on the economy, and having more educated productive people is a huge benefit that passes onto subsequent generations. As such government can justify giving away a hefty portion of tax revenue derived from people who would otherwise had remained poor had they not been successfully educated.

As Palacios says:

Education has great benefits for the student as well as for society. The measured returns to investments in education reflect under-investment in this activity, resulting in costs for individuals and for society. [This] under-investment in education is the result of a market failure.

In other words, there is untapped value in funding education. By making repayments contingent on income, this market failure is addressed because it allows investors to be fairly compensated for their risk by sharing in the upside. In the same way that equity finance powers business, so equity-type finance can power education.

Although paying a premium return can go some way towards helping to attract investors, there still needs to be a way to give funders confidence that the learners will actually succeed. If Government is struggling to educate a large portion of the population, why should we think that private investment would yield better outcomes?

Market forces can help achieve better outcomes

This gets to the heart of what makes GISA’s special and interesting.

Delivering a good education to millions of people is a massively onerous task for any government. Especially in a fast-changing world where it is hard to know what skills are needed for what jobs.

The South Africa government has recently committed to providing universal free higher education. But how efficient will it be in ensuring that the right courses and colleges are funded? One suspects that the amount of wastage and inefficiency will be massive. I do not think many people in South Africa believe that this measure will help deal with educational inequality.

I wonder if a GISA type program would be more effective because it passes on the complexity to the market. It introduces a self-adjusting mechanism that naturally optimises to deliver better education outcomes. GISAs increase the likelihood of education success because investors are obliged to make good decisions about where to put their capital. They must choose whether to fund blue-collar skills or white-collar skills, short-courses, or long degrees. They have to get these choices right or they do not get paid. This ensures that funds flow efficiently to where good educationb outcomes are expected.

Why should investors have better knowledge than government? Because experience shows that this is just how the world works. Investors who make bad predictions go out of business. Investors who make good predictions succeed and have more capital to invest. The market, for all its flaws, is better than government at creating value out of complexity.

There are other government funding models that incorporate market feedback such Social Impact Bonds3 and Income Contingent Government Loans4 (as offered by the Australian Government). GISA are theoretically more efficient than these because funds are distributed by private investors who are incentivised to achieve the best outcomes.

The value of market forces is that they could help accelerate the evolution of the education system by channelling capital to the colleges that society and learners really need. Public education today is more or less the same as it was 60 years ago. It is inefficient, ineffective, and ill-suited for the information age. GISAs presents an exciting mechanism to accelerate the reform of the tertiary education system to make it more efficient and relevant for the modern world.


Despite my enthusiasm for GISAs, I’m aware that the idea contains many open questions. How do you get government buy-in? What percentage of tax revenue per learner is paid back to the funder and for how many years do we repay? What is the repayment cap? Should we allow funding of traditional colleges only, or also allow funding for alternative education paths such as coding bootcamps and online courses? How do you ensure you’re not giving away too much tax revenue? How do you track the students and their tax payments and ensure repayment to investors? How do you achieve good results in higher education if the schooling is bad at lower levels?

It is beyond the scope of this article to cover these in detail. If this gets any attention then I will certainly look to explore the topic in much more depth. But in the interim I recommend reading the source of the idea - Investing In Human Capital by Miguel Palacios. It addresses most of the above questions and more.

I’ve only spoken about tertiary education so far, but there is only so much you can fix at this level of the education pipeline. So much of learning outcomes depends on what happens at school and even pre-school level (never mind factors outside of school that impact on learning outcomes).

As I mentioned earlier, Miguel Palacios put forward GISAs as a means to fund normal school education. The reason I focused on higher education is that it makes it much easier to test the model and establish its effectiveness, because there are only a few years between investment and job. But in theory, the GISA model could gradually be expanded to lower levels of school, and that is where it could really unlock much bigger benefits for society.

We know that investing in education is a vital building block for a more prosperous and equitable society. But we are constrained by available funds and by the complexity of delivering education equitably to the masses. In countries like South Africa, there are little signs of improvement in educational outcomes. Real reform needs to at least be considered. Government Income Share Agreement put forward one such vision.


Thank you to Miguel Palacios not only for originating the idea of Government Human Capital Contracts, but also for graciously giving of his time to discuss his ideas with me. And to Gilbert Pooley (co-founder of Umuzi) for reading my early draft and helping to stress-test the idea.


  1. The original idea presented by Palacios looks at ways to fund normal (primary and high) schooling, but for reasons that will become clear later, I’m going to discuss its application specifically to higher (tertiary) education) 

  2. For the record, Palacios actually called it “State Human Capital Contracts” 

  3. Social impact bonds differ to GISAs in that, with social impact bonds the repayments are related to more abstract measures than taxable income. For example, repayment may be awarded based on how many people find employment by X date. Since repayment is not tied into actual tax revenue collected, government needs to be careful not to award more Social Impact Bonds than what it can afford to pay. This limits the amount of potential investment in can accept. It also means that the effectiveness of the program depends on government chosing the payment metrics that actually add value. This is not as efficient as the market mechanism proposed by GISAs and it opens the door for sub-optimal outcomes to be achieved if the metrics aren’t well thought-out. GISAs are thus arguably more scalable, efficient and reliable. 

  4. Income contingent government loans differ to GISAs because in the case of the loans it is government, not the market, that decides who to fund, so it lacks the critical market mechanisms. However, in theory one could invite private investors to fund income contingent loans with government to repay these. Such a model would be comparable to GISAs. Having said that, income contingent government loans are still a move in the right direction as it incentivses government to deploy funds smartly in order to ensure they get repaid. One can see the affect of this with Australia’s recent decision to reduce funding for people who study humanity degrees.