Friday, November 14, 2008

Don’t Blame Charters

One of the accusations you'll hear our district leadership make is that charter schools are a drain on our school funding resources, and therefore one of the causes for additional operating levies.
This is a red herring. The truth is that the traditional education community just doesn't like charter schools because it rightly perceives them as competition. And they don't like competition.
Traditional public school districts, such as Hilliard City Schools, is funded from three sources: a) the property taxes (and sometimes income taxes) paid by the residents of the school district; b) the property taxes paid by the commercial entities in the school district; and, c) the pool of money redistributed by the State of Ohio.
The State's share is determined by a formula which can be observed via the SF-3 reports published by the Ohio Board of Education. Here is the most recent SF-3 for our district. The calculation goes like this:
  • Start with the the "Average Daily Enrollment," which is 14,333.73 (line 3). Multiply that number by the standard per-pupil amount that the Governor and General Assembly have agreed to set as the "Foundation Amount." For this school year, that number is set to $5,732. So 14,333.73 times $5,732 is $82,160,940.36 (line 5A). The theory is that we should be able give our kids the "thorough and efficient" education required by the Ohio Constitution by spending only this amount (we spend 1.75 times this amount).
  • The school funding laws require that each local district make a contribution to this cost via property taxes of at least 23 mills on our own local property owners. This amount, called the "Charge Off," is deducted from the first calculation. Our county auditor reports that the value of the property in the Hilliard School District is $2,419,400,820.86 (line 4C), making 23 mills be worth $55,646,218.66 (line 4E).
  • That which remains after the Charge Off is deducted is called "Formula Aid." It works out to about $27 million after some other adjustments (Line 6).
  • Then another set of adjustments are applied for things like transportation costs (line 16), Special Education (line 11), Gifted Students (line 14). We also currently have an adjustment associated with the elimination of Personal Property Taxes on businesses. It nets out to total State Aid of $37 million and change, or $2,628 per kid (out of our total per student spending of $10,665).
So how do charter schools fit into this?
The obvious effect is that our school district is required to send some of this state money to the charter school whenever a kid transfers. The SF-3 shows that this amount is about $1.6 million dollars for 233 kids (line 22B). That works out to $6,758 per kid. Here's the thing to remember: That's all the charter school gets - $6,758 per kid. From that, the charter school has to pay all of their costs, both operating and housing.
So what is the monetary effect on the public school district when kids transfer to a charter school?
Our district leaders want you to hear that the State gives us $2,628 per kid, and we have to turn around and give $6,758 per kid to the charter schools. They want to make it sound like the district is 'losing money' on the kids that go to charter schools.
But remember that we spend $10,665 per kid. The difference between the $2,628 we get from the state and that $10,665 is the revenue we collect locally via property taxes on residents and businesses. So when we send $6,758 to the charter schools, the district keeps the difference between that amount and $10,665, or $3,907.
"But wait," the school leaders will say, "you don't understand that when one kid transfers to a charter school, our costs don't really go down."
Actually, I understand this phenomenon very well, and have written about it before. Fixed Costs vs Variable Costs, or Marginal Cost vs Loaded Cost. The problem is that the school leaders want to use marginal cost analysis when talking to us about growth (i.e. more kids mean more cost), and fixed cost analysis when talking about a reduction in student population (i.e. costs don't go down just because there are fewer kids). Pick one or the other for both cases, please.
Whichever kind of analysis you use, the one thing that is clear is that the district is not harmed financially when a kid chooses to transfer to a charter school. I'd argue that the district is actually better off.
That is unless a lot of kids decide to transfer to charter schools. Then the district ends up with teachers with no kids and empty buildings. That would be a problem.
But the question you have to ask is why would that happen?
When Dick Hammond ran for re-election to the school board in 2008, the Hilliard teachers' union refused to endorse him because he sat on the Board of Directors of a charter school. Dick was unseated by Dave Lundregan, who did have the HEA's endorsement. During the candidate's night, we four candidates were asked if we supported charter schools. Mr. Hammond's answer was spot-on: If the public school district was doing its job for those 233 kids, there would be no need for the charter schools.
Administrators and teachers think charter schools are bad because they're bad for teachers and administrators, not the kids who attend the charters.
But it's all about the kids…
This article was motivated by a recent report published by the Buckeye Institute


  1. Paul,

    You went through the whole SF3 calculation but missed the essential point: Before the charge-off is met, the state funds the marginal student at $0. After the charge-off is met, the state funds the marginal student at the full basic aid amount: $5732.

    You fell for the trap set by school districts -- the traps that says the state funds the marginal student at the state share. That is absolutely false.

    Of course, there are categorical aid amounts involved that are funded per pupil or include pupil demographics in the calculation.

    So, the marginal student adds $5732 to SF3 funding when they enter a public school and take only the $5732 when they leave (assuming the charge-off is met). The categorical aid amounts (special ed, etc.)are where a district takes a hit since here the state share IS part of the calculation.

    At the margin, a charter student is no different from a student who leaves the district for any reason, including moving out of the district, attending private school or home school, or taking advantage of inter-district enrollment (if offered). The net in state aid is zero (excluding the categoricals).

    If a district planned correctly, the marginal student who is destined for a charter added no expense whatsoever. All fixed operating costs are associated with the annual budget (buildings are a separate issue, and are, or course, funded via bond issues. There are some long-term leases that span school years, etc, but these are the exception).

    A district that accurately predicts its student population does not suffer due to charters (oh, sure, teachers may lose their jobs as in Columbus, but that is a separate issue from the district's financial position). A district that plans for too many students sees a hit on their finances -- less aid than predicted but fixed costs that cannot be reduced during the school year. But this is true regardless of why student enrollment was less than predicted.

    A district that plans for less students sees a financial bonus -- more state aid with likely minimal added costs (Paper, water, etc., since districts usually do not add sections during the school year).

    All that said, Hilliard is on the guarantee so the whole per-pupil funding issue only works to set the guarantee amount.

    So, it is true that your district now gets zero for the marginal student yet loses the full amount. But the fact that your district is on the guarantee means that it is being funded above the calculated state aid -- in essence, Hilliard is receiving a bonus year after year.

    note: Olentangy will likely remain in the same boat after the October ADM.

  2. Mr. Fedako is correct - I missed an important nuance in the SF3 calculation:

    There is a calculation on SF3 at line 19 called the "Transitional Aid Guarantee." It serves to maintain year-to-year funding at the same level even as other components decrease. The effect is that when a marginal student is added to (or subtracted from) enrollment, the guarantee is decreased (increased) by a corresponding amount, and the net effect is zero.

    However, the payout to the charter school is accounted for below this guarantee line, meaning that the full amount transfered to the charter school is deducted from the state funding. The actual amount transfered is affected by a few items, notably the special education category the transferee fits into, so the actual amount transfered may be the whole $5,732/pupil of basic state aid, plus additional funds associated with special ed supplements. On average, it's the $6,758 I reported above.

    So what's that mean taken together? It means that when a kid transfers to a charter school, the school district loses $6,758 in real revenue, but probably has no change in cost. This is the marginal analysis, and a critical perspective on the situation.

    But let's also look at it from the fully-loaded basis.

    Assuming revenue equals expenditure, the fully-loaded revenue per student is $10,665. Therefore if the payout to charter schools average $6,753 per student, then $3,912 in revenue becomes unencumbered and can be allocated to increase the per student spending on the students remaining in the district schools.

    Neither form of analysis is the perfect one to use in all circumstances. For example, what is the cost per student for a $65 million high school designed to house 1,800 students?

    The fully-loaded answer is easy: $36,111.

    However, on a marginal basis, the number is zero. If student number 1,801 shows up, the building can handle it as is. How about student #1,802? Still can handle it.

    So how about student #2,341? Yep, that's the one that puts us over. That student costs us another $65 million, then student #2,342 is free again. That's not a helpful way to look at our facility costs, is it? The fully-loaded approach gives more insight when analyzing macro conditions - ie how much should we expect to spend if we add another 3,000 high school students to the district? It should be in the ballpark of $36,000 times 3,000, or around $100 million (yikes!). Not 3,000 times the zero we come up with in the marginal analysis.

    On a fully-loaded basis, we take in more revenue per student than we pay out to the charter schools. For purposes of setting macro-level public policy (ie - answering the question: Do charter schools bring financial harm to public school districts?) the fully-loaded analysis is more appropriate.

    And the answer is No.

  3. Paul,

    To tighten up your points:

    You are conflating expense and debt. The $36,000 is additional debt, not expense. For capital-related expense figures, you need to go to your district's audit (found on the state auditor's website) and add interest paid to the capital-related charges. Then, to understand the true total expenditure amount, you need to add in any other expenses not already listed as an expense to the general fund. Finally, you have to decide which ADM to use in order to get your per-pupil expense.

    Be cautious that you do not double or triple count as there are three figures you could use for capital-related charges: 1. depreciation; 2. actual capital outlay for the year from the bond fund; 3. principal retirement from the debt service fund. Make certain that you include OSFC moneys if applicable.

    Depreciation is the best since it allocates expenses over the assumed life of a building. Actual capital outlay is not valid since it charges the full amount in one year. Principal retirement is not valid since some districts -- such as Olentangy -- push debt out in the future (Olentangy had one bond issue for $65 million where only $1,000,000 was paid in the first 10 years [$100,000 per year] of a 27-year bond issue.

    Also, make certain that you include other funds since the federal title programs may not be included in the numbers that you are using -- again, the audit has all of this.

    I always tell folks to go to the audit though I had a superintendent state that I was lying by using those figures. Lying by using the district's own audited numbers!?! Never quite understood that one -- though it was a good PR move by the superintendent.

  4. Jim:

    My writing could perhaps have been more clear, but I have no trouble understanding the concepts of capital cost, operating expenses, depreciation and debt service. My degree is in business, and I got "A"s in all my accounting and finance classes. If you reread my comment, you'll see that I was addressing the notion of marginal vs fully-loaded analysis by using the cost of a building as an example.

    Clearly there are a lot of moving parts in school funding, and it's easy to let all the 'noise' distract from the macro issues.

    For us here in Hilliard, and in other suburban districts as well I suspect, there are really only two issues: a) suburban municipal governments have let residential developers plunk down hordes of houses without a corresponding amount of commercial development; and, b) employee salaries and benefits are growing faster than the valuation of the real estate on which the revenue is based (in fact, they're going in opposite directions right now).

    On top of that, the State of Ohio has an increasingly serious budget problem, and are unlikely to fix that by coming to the rescue of the 'wealthy' suburban districts.

    This charter school stuff is mostly noise. But if we let the public school folks keep making false claims without challenging them, it becomes accepted as truth - just like any propaganda.


  5. Paul,

    All A's aside, you are still conflating debt and cost. While this may seem like a wasted exercise in semantics, it is an important difference to understand.

    That said, back to your analysis: If it is true that you opened the new school for student 2,341 only, all costs associated with the new school must be place on that student. Why? Because you have to consider the cost of a new school against its alternatives (trailers, rented space, etc.).

    In business, if you have a machine that produces 1,000 widgets per year, and you considered buying machine two in order to produce widget 1,001 only, the yearly full cost of machine two would be placed on widget 1,001. You certainly wouldn't average the costs of the two machines over 1,001 widgets since that would give a false impression of the value machine two added to your firm -- and would lead to an incorrect buy/no-buy decision.

    Likely, the school is opened with the plan to house students for 50+ years. Likely, at least 541 students would be in the school for the next 50 years (2341 - 1800). Or, at least, that is how the analysis should be performed.

    Otherwise, your district should simply rent a trailer to house the singular student for his/her four years.

    Both Hilliard and Olentangy need to go through such an analysis before student growth peaks and then falls. Typically, a district can see a 10% student population drop after build-out, when the population matures. So, opening the last school may lead to boarded-up, relatively new school ala Worthington.

    And then the cost of the building cannot be spread over any student. The building is just an empty money hole.

    Our board and voters just committed to expanding a high school now, based on a 10-year projection, even though the added space may never be needed.

  6. Jim: Trust me, I understand - I'm just trying to illustrate the difference between fully-loaded and marginal cost analysis.