Operational Efficiency #7: Ten Tentative Conclusions

Based on my previous posts, here are ten tentative conclusions that I've drawn:

1. It is important to think critically about operational efficiency, particularly the way the data is presented and the conclusions drawn from it. Board members and citizens have, in some cases, drawn erroneously conclusions based on a misunderstanding of what the data indicates.

2. Bigger schools do tend to do better in terms of operational costs than smaller schools, but there are a significant number of exceptions particularly as schools reach 300 students or if smaller schools used mixed-grade classrooms effectively.

3. We need to pay better attention to when new administrative help is needed to run a larger school. There are some elementary schools where an Assistant Principal is hired, and more that have School Administrative Managers. At whatever point we decide that this extra help is needed, it will cut against the operational efficiency of a school. Financially, it may make more sense to stop just before reaching this point. It would be helpful if there was greater clarity about when the decision to add more school administrative help occurs and on what basis. I think this is a good reason that operational efficiency is more mixed between 300-student schools and 400+-student schools.

4. If a school has at least 300 students, the number of exceptions to (2) grow a great deal. The likely reason is that if a school has 300+ students, then it likely has a sufficient number to fill two classes in every grade in a sustainable way. Nevertheless, the cost of having 300 person schools is that class sizes will sometimes be lower and sometimes be higher than we might prefer, and having these types of schools likely means that the district shouldn't respond unless the class sizes are grossly out of line.

5. If a school has fewer than 300 students, then it probably makes the most sense financially (a) to share some administrative and support positions with other smaller schools and (b) to have mixed-grade classrooms (with appropriate training and support for conducting those classes well). If mixed-grade schools are unacceptable for academic reasons, then that is a significant mark against having schools with fewer than 300 students.

6. We should control for the use of special education and at-risk funds. The "total cost to operate" a school can be useful, but it is seriously and egregiously misleading when it is compared to the per student general education allocation. Furthermore, the use of special education and at-risk funds is largely a function of board and administrative decisions (e.g., the location of Hoover's autism special education classes), and it is deceptive to compare the overall operational costs at such schools with schools that need few resources in those areas. There are various ways to control for these costs, as I discussed previously, and we should pay attention to how its done. We also need to make it clear to the administration and the board that this needs to happen when it's not done.

7. We need to be skeptical of the ability of new schools to offer a savings, even over some of our least efficient schools, at least in the first five years of the school's existence. Sometimes the total cost per child information has been used to make it seem as if they do better than they do. But that has been because of a low proportion of special education and at-risk funding at the school.

8. The $500K figure for opening a new school is not the amount that would be saved by closing a school. That figure depends on the relative operational cost of a school and other factors. The $500K figure is useful because it tells us that we need to prepare for our newer, likely to be inefficient, schools.

9. The operational efficiency of a school depends on much more than simply its size. The demographics, income levels, administrative decisions, and so forth play a much greater role than is often presented to the public.

10. Before proposing the closure of a school on operational efficiency grounds, we need to consider ways to have it become efficient (which doesn't just mean expanding it.).

Prior Posts in the Series:

Part 1: Introduction
Part 2: Comparing Seven Schools
Part 3: Comparing 2 of our smallest schools to our newest, larger school
Part 4: A deeper discussion of the rationale for controlling for special education and at-risk funding
Part 5: A defense of keeping Hills
Part 6: A critical evaluation of the $500K figure and it use

Operational Efficiency #6: Won't we save $500K by closing a school?

Over the next couple of days, you very well may hear ICCSD administrators or school board directors assert that it will take $500,000 to operate our new elementary schools beyond the instructional costs. This figure is based on the projected salary of a principal, librarian, guidance counselor, building secretary, media secretary, custodian care, and utilities cost. The estimated new cost to the district is around a half a million dollars. The way that this figure ought to be used is to address the inefficiencies of our new schools when they come online (see my earlier post comparing our smallest schools to our newest, bigger school for some evidence of this fact). That is, since they will likely hamper our ability to be more operationally efficient in the short-term, it is a good idea to accumulate some unspent spending authority sufficient to, at least, cover the new cost to the district of operating a new elementary school.

As I mentioned last post, spending authority is the amount the district is allowed to spend in a given year. If the district does not spend the full amount, they are allowed to carry over the difference between their total spending authority and how much they actually spent as a one-time rainy-day type fund. We will need those extra funds when our new schools come online, and it is a good idea to prepare for it in the coming year(s).

Some, however, go well beyond this use of the $500K figure. They argue that, since each elementary school costs approximately $500K to operate beyond instructional costs, that we'd save $500K just by closing schools and consolidating those students into other school. I have heard this argument used by school board directors and local citizens, but it simply doesn't bear up under scrutiny.

I've heard that argument applied to Hills. It is said that if we just closed Hills and moved those students to another school that we'd save $500K. Unfortunately, Hills has only $250K that it spends on non-instructional expenses. Some of those -- media, custodial care, and utilities costs -- would likely be needed at other schools to accommodates those students were Hills closed. With regard to Hills, the $500K figure is actually more like $200K. 

What about another school? The $500K figure has been tossed out with regard to Hoover's planned closure too. In fact, the operational cost saving was presented as one of the justifications for closing Hoover. But the figure doesn't apply to Hoover either. In order to show this with the 2012-2013 data, I assumed that the 364 Hoover students were distributed to the surrounding schools -- 100 to Longfellow, 100 to Lemme, 100 to Lucas, and 64 to Mann. Doing this would save the district money on by eliminating the need for Hoover's administration and most of its media services (staff) but it probably wouldn't significantly affect utilities or custodial costs. The bigger schools that would be needed to accommodate the Hoover students would probably make utility and custodial costs a wash. Under this scenario, I examined how much the district would probably save. The total amount spent on administration and media services at Hoover in 2012-2013 was approximately $275K. So, one might think that the district could save that much by closing Hoover, but that wouldn't quite be accurate either since Hoover is significantly better in terms of operational efficiency than 3 of the 4 schools that its students would be moved to (Longfellow is the exception). Thus, by my calculations and assuming that every Hoover student was moved to one of these other schools, the district would save approximately $191K. Of course, if a few students opted to switch to another neighboring district (e.g., West Branch), a private school, or home school because of the change, then the amount saved goes down even more. 

So, to make a long story short: use the $500K figure the way it was intended. It doesn't tell us how much we'd save by closing a school. It tells us what we need to save to pay for our (likely to be inefficient in the short-term) new schools.

Prior Posts in the Series:

Part 1: Introduction
Part 2: Comparing Seven Schools
Part 3: Comparing 2 of our smallest schools to our newest, larger school
Part 4: A deeper discussion of the rationale for controlling for special education and at-risk funding
Part 5: A defense of keeping Hills

Operational Efficiency #5: But shouldn't we close Hills for the sake of operational efficiency?

Given the Iowa City Community School District (ICCSD) is concerned about budget shortfalls and saying that "Nothing is off the table" with regard to budget cuts, I've been discussing how the ICCSD calculates, analyzes, and evaluates operational costs. This post is the fifth in my series; links and short descriptions of prior posts are at the bottom.

Numerous school board members have previously expressed how costly it is to run Hills Elementary. It is the smallest elementary school in the ICCSD  at an enrollment of 101 students for the 2013-2014 school year, and it had an enrollment of 108 students in the year I have been examining (2012-2013). Furthermore, as many may have surmised, no matter how the numbers are parsed, Hills currently has the worst operational efficiency in the district. It is also something of an outlier compared to every other school in the district. 

Still, the way the district ordinarily presents the operational costs is misleading. Likely because of generational poverty within the boundaries of Hills, it has, proportionally, some of the highest special education and at-risk funding in the ICCSD. So, if you hear district administrators' say that it costs  $10,000+ per student, they aren't controlling for these funds as I've suggested they should.

Nevertheless, that doesn't mean that Hills is operationally efficient by any measure. When I control for these costs and evaluate the amount of funds provided for ordinary school operations from the general education budget (see my prior post for a discussion of what I mean by this), Hills, at best, has a per student operational cost of $5,964. It is slightly more than $1,000 per student compared to the next most expensive elementary school in the ICCSD.

Based on this information, the operational efficiency argument for closing Hills seems well supported. And that is the conventional wisdom. 

There are a few things that still concern me, however. First, the $5,964 figure is still slightly below the $6,018 allocated for each student in the 2012-2013 school year. In other words, it is Hills' need for special education and at-risk funds that results in it "losing" money.

Second, it could very well be the case that closing Hills Elementary would result in the district losing significantly more money than it loses by operating Hills Elementary. How so? In the 2013-2014 school year, the ICCSD is losing 91, 52, and 10 students to the Mid-Prairie, Highland, and Lone Tree school district's respectively. That's a total of  153 students, and losing those students costs the district over $900K a year. I'm worried that we could see a significant number of the Hills community follow this trend taking their kids to other school districts. The only way closing Hills really saves money is if a sufficient number of those families decide to send their kids to another school in the ICCSD rather than sending them to another school district. Otherwise, closing Hills could very well result in a net loss to the district.

Third, the ICCSD has not sufficiently attempted alternatives to closing Hills. My previous post comparing two of our other small schools, which are well over a 100 students larger than Hills, suggest that if Hills had another 100 students, it could probably be run, especially with combined-grade classes at least as efficiently as some of our other schools in the district. The district should consider ways to make that happen. I wonder what a public Montessori, or a progressive public charter, like this one might do to help Hills. I'm not sure if it would improve the situation, but it important to consider these possibilities and to try something before making a decision that could have significant unintended consequences. An attractive Hills Elementary could be an effective way to draw back a number of families who have left the ICCSD for those other school districts.

Fourth, this post has only been an evaluation of the operational efficiency argument for closing Hills. It is an important one, but there could be academic and social reasons for keeping it open that trump this argument. 

Previous Posts:

Part 1: Introduction
Part 2: Comparing Seven Schools
Part 3: Comparing 2 of our smallest schools to our newest, larger school
Part 4: A deeper discussion of the rationale for controlling for special education and at-risk funding

Operational Cost #4: How to control for special education and at-risk funding?

I've been doing a series on operational costs in the Iowa City Community School District  (ICCSD). Part 1 gave an introduction to my methodology and general approach, Part 2 discussed 7 of the ICCSD elementary schools arguing that 4 smaller ones were significantly more efficient than the larger ones, and Part 3 compared the newest school with two of the smallest schools in the ICCSD -- and the newer school didn't fare so well.

A concerned reader has argued that I am distorting the numbers. The problem, according to this reader, was that I was inappropriately controlling for special education and at-risk funding. As I argued previously, there are three separate funding categories for general education, special education, and at-risk funding. The state allocates a certain amount per student for general education purposes (that figure was $6,018 for the 2012-2013 school year), a different amount for special education depending on the needs of the students in the district, and another amount for at-risk funding (which is only funded at 75%, so 25% comes from the general education budget). The reason I wanted to control for special education and at-risk funding is that it is misleading to compare the $6,018 figure to how much a school costs to operate in general. It doesn't take into account the sources of those funds.

Why is it important to take into account the source of funds? The primary challenge our district has financially is not the ability to raise funds. Rather, it is a problem in being allowed by the State of Iowa to raise funds. In other words, the primary limitation is related to spending authority. The idea behind spending authority is that it ensures that all students in the State of Iowa basically get the same type of education -- and wealthier areas don't get more funds simply because they are from a wealthier area. And that's a good thing.

For the most part, ICCSD's special education funds and 75% of its at-risk funds come from sources that don't count against the spending authority related to the $6,018 figure. Furthermore, they are not a true measure of how costly it is to operate a school. Rather, they are a measure of how costly it is to operate that school given certain decisions about special education and at-risk funding that the district has made.

But why should we control for these funds? It turns out that special education and at-risk funds aren't equally distributed at all in our district. I know what you are thinking! Surprise, surprise! So, to really compare our schools apples to apples, we need to control in some way for this variance.

Now, my well-informed and well-meaning reader was concerned with my method of controlling for these other funding sources. In short, I had zeroed out special education funds and at-risk funds for all 19 of our elementary schools. This concerned reader argued that this isn't the best way to control for those sources and it advantages the smaller schools. I thought there was something to my interlocutor's point, so I crunched the numbers a bit more.

In my first attempt, I took the mean of all 19 elementary schools special education and at-risk funds, and I put that amount in for each school. Under this way of controlling for those external funding sources, larger schools tended to fare better. They had a larger base to distribute that average over and they did quite well because of that. Here is the graphical representation of the data:

A few notes: first, we need to be clear that the cost per student here is a way of representing the total cost to run a school and it should not be compared to the $6,018. So, it loses some of what warrants the zeroing out strategy I used previously. Second, this approach is arguably problematic because it acts as if special education and at-risk costs should be distributed evenly across the district. But doing that would result in smaller schools having a disproportionate number of students with special education or at-risk needs.

Thus, I came up with another way to control for the external funding sources. I took each of our 19 elementary schools enrollment as a percentage of total enrollment, and I allocated exactly that percentage of the total elementary special education and at-risk funds to each school respectively. Here is the graph for that data:

As one would expect, the result was representationally, virtually identical to my initial portrayal in Parts 1-3. So, you don't have to click back, here is that graph:


In other words, if you allocated special education funds and at-risk funds in our district according to enrollment, the comparative operational costs rankings for all of our schools would be all but identical. I think there are good reasons for (a) distributing those funds by school enrollment rather than acting as if the best way to control for those costs is by equally dividing it, and (b) there is a good reason for zeroing them out, particularly since it doesn't skew any comparative evaluation of our schools and the benefit is that it gives us a pretty good comparison of how much each elementary school costs our general education fund. The only caveats that I should make here are twofold: first, at-risk funding does affect the general education fund and it is a bit misleading to not include 25% of it here. Nevertheless, we still need to find some way to control for that with something like my second graph. Furthermore, our district has a so-called special education deficit. That means that we spend more in special education than we receive from the state. So, I could see how it would be helpful to distribute those excess funds across all schools as well. But once again, that graph would give the same comparative rank of our schools as the 2nd and 3rd graphs above.

In short, I thank the concerned reader for bringing a possible oversight to my attention; but my considered reflection is that it my initial portrayal along with my 2nd graph above (understanding the limitations of both) gives an accurate picture of the comparative operational costs of our 19 elementary schools in 2012-2013.