# Let’s Design a School, Part 3.2 Costs

We’ve established our hybrid model of school, where social services issues are addressed with a sieve model and education boils down to three phases.

In part 1, we went through the expected costs for the school-as-social-services model.

Now we’ll go through the costs for school-as-education, then we’ll discuss the costs of the full model.

Now let’s talk numbers.

# Costs of School as Education

We earlier estimated that, for an example school of 100 children with a budget of \$2 million per year, roughly \$500k would be spent on non-employee costs (electricity, food, taxes, transportation, etc.), leaving us with \$1.5 million to spend on employees. (We’re ignoring capex costs right now, e.g. the cost of building a school.) This \$1.5 million was then scaled down to \$1.125 million to account for the non-salary costs of each employee (if an employee’s salary is \$75k per year, their total cost including paid time off, health insurance, etc. is closer to \$100k per year).

So how do we allocate this money?

First of all, because a great deal of the curriculum will be handled online, each student will need access to a computer.

Taking \$1,000 per laptop per student as a general rule, this costs \$100k total. On the other hand, students don’t need a new laptop every year, just access to one every year. So while the \$100k won’t cover the costs of access to a computer for all students for all time, it isn’t something we have to pay every year either.

If we assume students are about evenly spread out age-wise, then we’ll only need about ten new laptops per year. Double that due to damage, theft, or other loss, and we’ll budget \$20,000 for twenty new laptops per hundred students per year. Maybe add another \$5k per year for peripherals or other expenses.

This brings our total salary bank down to \$1.475 million/​year.

## Teacher Salaries

The median American teacher salary is somewhere in the vicinity of \$66,000 per year. While the variance on teacher salary can be large, we’ll go ahead and use a salary of \$70k per year for our teachers in this hypothetical school.

If our budget was entirely spent on teachers, that would net us 15 teachers for 100 students, or about one teacher per seven students, which is an incredible ratio.

That being said, I don’t think we need that many teachers.

Remember, the entire curriculum for every class is already designed and available online, as well lectures, exercises, etc. All the survey courses and specializations students go through in phases 2 and 3 are either online or done through work-study or apprenticeships.

In other words, the school needs teachers for in-person learning, answering questions, and guiding students academically, but not much else. The teachers we do have will thus be free to spend their time actually helping students, putting together various activities or clubs, tutoring, or organizing in-person seminars.

And we don’t need one teacher per seven students to do that.

## Capex

The term capex refers to ‘capital expenditure’, and means the initial investments in infrastructure, physical or otherwise, needed to get a project up and running. So far we’ve been ignoring capex in our model, especially because the school buildings and infrastructure already exists.

We will need, however, to create the online curricula that schools will be using to teach the core requirements of phases 2 and 3 (civilizational skills and adulting skills, respectively). The way I envision it, this would largely be the responsibility of the federal Department of Education, and thus wouldn’t involve the budget of individual schools.

What each school may need to pay for, however, is access to various online learning platforms containing the survey courses and specialized courses students will have access to. A large amount of this is currently available online for free, but at the scale we’re talking about it makes sense for some kind of licensing arrangement.

Coursera, a popular platform for MOOCs (massive open online courses), costs \$400 per user per year for the business plan (5-125 people). While the price for larger groups isn’t listed, it would certainly be cheaper.

Because we are talking about millions of children and the US Federal Government, assume that the contracts for access to these educational resources are done at massive scale, driving the price down further.

Just to pick a number, assume that access to a variety of educational resources costs \$1,000 per student per year, or \$100k per year in our example school. For the cost of a single teacher’s salary per school, students are given free access to almost every educational resource on the internet.

# Costs of the Full Model

Remember that we estimated \$500k for non-salary costs in part 1. From that we’ll add \$100k for access to online learning, bringing our total budget for non-salary expenses from \$2 million to \$1.4 million. Including computers brings us down to \$1.375 million.

Because the full cost of an employee is more than their salary, we multiplied our total salary budget by 34 to come up with a total budget for base salaries.

This leaves us with \$1,031,250 for salaries.

The salaries for the various roles we’re using:

• Social workers will cost us \$75,000 per year

• Counselors will cost us \$60,000 per year

• Nurses will cost us \$60,000 per year

• Adult supervisors will cost us \$40,000 per year (This isn’t exactly an existing job, and should require little qualification beyond background checks.)

• Teachers will cost us \$70,000 per year, as per above

• Administrators will cost \$100,000 per year

So what can we do?

One of the great strengths of this kind of school is the flexibility for each school to allocate its budget towards more of what its students need. That being said, each school, no matter how extreme its needs, will have at least one of each role.

Adding that up, that’s \$405,000 out of our \$1,050,000, leaving us with \$645,000 to spend on additional educational or social services resources.

To illustrate how different schools may budget, we’ll look at two cases: a school that’s mostly education, and a school that’s mostly social services.

Mostly educational

• While technically the school could afford an additional nine teachers (for a total of ten), that seems a bit too extreme. Instead we’ll trade two teachers for an additional counselor and social worker, leaving us with:
2 Social Workers
2 Counselors
1 Nurse
8 Teachers
For a total spend of \$1,030,000.

• This is 15 adults per 100 students, with 8 teachers per hundred students, or ~12 students per teacher, which is smaller than the American average of ~16.

• Such a school should be able to handle the occasional problems its students will have, but is primarily focused on educating them.

Mostly social services

• We’ll cap our teachers at 3 - remember, the teachers are there to interact with students and help them, not design curricula or lecture. Furthermore, out of the 100 students served, there may not be more than a third who are in a good state to learn on any given day, meaning we have the same number of learning students per teacher as the mostly educational school. This leaves us with:
3 Social Workers
3 Counselors
2 Nurses
3 Teachers
For a total spend of \$1,035,000. (Close enough, given the kind of estimation we’re doing.)

• This is 17 adults per 100 students, or about one adult per five children. Each social worker/​counselor would be responsible for 33 students, which is roughly enough for them to see each student for one hour a week. Not ideal, but far better than existing schools.

• In extremis, the number of teachers could be cut down to two or one if the school needs more resources devoted to social services resources.

In the end, this sort of estimation is never supposed to be a knock-down argument. What it is is an existence proof—it seems financially possible to create the kind of school that I’ve written about, and staff it appropriately.

# Conclusion

It seems possible to create a school based on the structure and curriculum we’ve outlined, given a reasonable budget per student. At the very least, a rough calculation shows that our school is plausible.

Given the details described so far about our new kind of school, we’d like to open up the comments for questions, criticisms, or other feedback on the model that we might address in future posts. We’d like to solicit a wide variety of viewpoints to better understand how this model of school might be better (or worse) than the existing public school system.

So please, feel free to chime in!