Pricing Education: A Worksheet

Not too many years ago, someone respectable described to me K-12 education in the United States as "glorified daycare" and regardless of my own opinions on the matter, the concept stuck in my mind and has been bothering me ever since.

A LinkedIn post on the unaffordability of childcare (again, in the United States) finally made me look up and work out some numbers... In what follows, I assume that we are looking to cover a child's complete upbringing, starting from daycare through high school (age 18); I am looking to find the discretionary cost of bringing up a newborn into an adult, that is the incremental cost that would allow parents to have a full-time job without sacrificing the quality of their child's upbringing.

Bottom-Up Calculations

Starting at birth, for a single child:

• the first three years require a full-time adult per three children (cost: 1 full-time adult year, FTAY from here on)
• the next fifteen years require a full-time adult per fifteen children (cost: 1 FTAY)
• assume that the above only cover "academics" and factor in the need for an adult paying close attention to smaller groups of children, maintaining familiarity and bonds through graduation: 18 years at a ratio of one adult per 18 children (or more realistically, 9-ish years at a 1:9 ratio) which incurs another 1 FTAY.

Total child bringup cost: 3 FTAY.

Let's assume one FTAY ranges from $80K to $160K: California average salary is around $90K, per CSBA, and that translates to $120K average yearly cost to the employer if we consider various benefits and taxes. We can conclude that child upbringing (0-18) that allows parents to have a full-time job and not feel like they are depriving their child has direct human cost of around $360K, or $20K/year/child. Note that we did not account for overhead (buildings, management, governance, etc.) which can be expected to easily bring the number into the $30-$40K range.

Top-Down Calculations

Consider reported California state spending per child: about $23K yearly per pupil for K-12 children according to the 2021-22 budget act (source). According to the same page, total spending is about $124B, which implies a student population of about 5.5M. There are two implications:

• A spending gap of about 50% ($62B) to bring the per-pupil budget numbers to that bottom-up estimates above.
• A further gap of 40% ($186B * 40% = $74B) to account for the "first five years" that are currently not covered by the K-12 scheme.

Next, the talent pipeline:

• CSU Sacramento (semi-randomly chosen for the example) reports about 8.5K graduates per year (dashboard), of whom 476 are getting a bacherlor's degree from the College of Education. This is a ratio of 1:18. If we count all degrees awarded (786) we get a much more beneficial (education-forward) ratio of 1:10.8.
• If we expect that, on average, every adult is "responsible" for one child, that means for every adult's 40 years of work there need to be 3 FTAY completed by an educator, for a ratio of 3:43 or 1:14.3. It appears that on the talent supply side, the numbers roughly match (assuming all graduates are passionate for the career they have chosen and pursue it for 40 years).

A Second Look at the First Years

Those who read the above carefully will notice that I cheated and bundled two separate issues together. In my model, the first three years contribute to the cost disproportionately (at least a whole 1 FTAY). One could argue that since the system is not trying to cover years 0-5, we should exclude them in order to have a more clear argument. This has two implications:

• There is no solution in sight for the early-years gap that we have, and reported high costs are "normal replacement costs" in the context of working parents trying to compensate for the years they are working full time will having a toddler at home who needs to get properly taken care of. Let's call this "the first years gap".
• There is, at least in theory, not much of a budget gap for the later years, which under the revised model should only cost 13/15 + 9/9 = 1.86 FTAY or approximately $17K yearly. This is within the current per pupil budget, allowing even for a modest $6K/year overhead. Given my experiences, there must a significant quality and/or overhead issue because I have not witnessed a level of service that would convince me that "the system is taking care of my children while I am working full time". Let's call it "the K-12 quality gap".

I suspect both gaps above are partly due to the highly varied expectations that the public has in a place such as California regarding spending, quality, and overall philosophy of child upbringing (in other words, my personal model for how this should work is irrelevant). There are countries where there is a more consistent, shared understanding about how children need to be educated (in the broadest sense of the word).

Opportunity for AI?

My intuition, observations, as well as the above numbers suggest that the gaps between my (subjective) "ideal" budget on the one hand, and the one afforded by the state on the other, are in practice bridged by stretched student to teacher ratios. For example, classrooms I have had direct visibility into lean towards a 1:25 ratio rather than the 1:15 we used above.

For the sake of argument, let's consider the current budget immutable (since spending more or less money is always a politically charged topic). I also do not anticipate that in the foreseeable future AI will excel at bonding with humans and being a steady, trustworthy presence that helps children orient themselves and mature as they grow up.

The opportunity at hand is to delegate the academic part of the upbringing process to machines: with strides being made in training AI models that exhibit reasoning capabilities, I anticipate that one out of the three FTAY budgeted earlier (the 15 students for 15 years academic teaching segment) can be automated; this in turn would make the overall current budget satisfactory as long as the automation has a low, amortized up-front cost and low marginal cost per student. The risk of course is that instead of leading to improvements in quality, a shift like this will be seen as a way to cut total state spending back down ??.