SFUSD's Mammoth Facilities Need
How a cash-strapped district can save $100 million
Later this month, SFUSD is supposed to produce a first draft of its Facilities Master Plan. This will lay out all the money that the district plans to spend on school site maintenance and upgrades over the next ten years. A proposition to pay for all this work is planned to be on the ballot in 2024.
Everyone accepts that the district’s facilities need to be upgraded. But twenty three (!) sites whose refurbishment was supposed to be paid for by the last Facilities Bond, in 2016, are still awaiting the promised work. It is reasonable for voters to have doubts over how prudent the district is about spending the money it is given and whether it will be able to deliver on any promises it makes to voters next time. The district wants to make a compelling case to voters.
The Facilities Condition Assessment
Before you can have a plan about what you want to do, you need to know what the current situation is. Last year, the district conducted a survey of all its sites and produced a Facilities Condition Assessment. Included in that assessment was an estimate of how much it would cost over a 5-year period to bring each school site up to the desired level. The total was an eye-popping $1.43 billion1. That’s more than the district’s annual budget. High as the estimates are, only a foolhardy optimist would assume they are too high.
Facilities Condition Index
The centerpiece of the Facilities Condition Assessment was the Facilities Condition Index (FCI). The FCI is the ratio of how much it would cost to repair and maintain the facilities over the next five years to how much it would cost to completely replace them. Intuitively, the higher the FCI the poorer maintained the facilities at the school are. Low FCI: good; high FCI: bad.
The range and distribution of FCI values is shown in Figure 1 below.
The school site in the poorest condition is the Mayeda campus of Rooftop ES, followed by Mission HS.
A natural question is whether schools in certain parts of the city are in better condition or more in need of repair than schools in other parts of the city. At the board meeting, Kamalanathan displayed a map (see figure 2) which shows the Facilities Condition Index (FCI) at each site.
Her observation was:
“here’s a map just showing how the sites that have been renovated and are in good shape are distributed across the city as well as the sites that are poor and deficient and overall we see pretty even distribution patterns with […] the exception of the Mission and in SOMA, where we see some clustering there”
Indeed, most of the schools in the Mission and SOMA are in yellow (“fair”) or orange (“poor”) condition and there are fewer in light green (“good”) or dark green (“excellent”) condition. But the boundary between, say, “fair” and “good” is arbitrary and it is possible to be more precise.
Elementary Schools are better maintained than middle and high schools
Figure 3 shows the average FCI by type of school. Elementary schools have a considerably lower FCI than middle and high schools, indicating that they are generally in better repair and require less work. We’ll see later how SFUSD’s policies and practices cause the larger schools to be neglected.
Latino middle and high-schoolers have the worst school conditions
We can calculate the average FCI experienced by students from a racial/ethnic group by weighting the FCI for each school site by the number of students from that racial/ethnic group at the school. This will show whether one group systematically experiences better or poorer conditions than another. Figure 4 shows FCI for the largest demographic groups in SFUSD.
At the elementary school level, the average FCI experienced by Asian students is 23% and by White students is 26% with the other groups between those values. The gap of 3% between those two values is half the size of the gap between elementary and high schools and so is not concerning to me. Incidentally, part of the reason that White students have the highest FCI values is that Grattan and New Traditions, two of the three schools that are majority White, have FCI scores of 0.38 and 0.37 which puts them firmly in the “poor” category.
At the middle school level, Latino students have a much worse average FCI (33%) than any other group and the gap between the highest and lowest is 8%, which is more concerning. The reason for the gap is easy to identify. Of the three middle schools in the poorest condition, two (Everett and Lick) have large Latino majorities and the third (Denman) has a Latino plurality.
Latino students also experience the worst high school conditions, on average, but the gap between the highest and the lowest is not as large as it is in middle school. The three high schools with the poorest conditions are SF International, Galileo, and Lincoln and these are three of the five high schools with the most Latino students. Galileo and Lincoln are the two schools with the most Black students, and they are both 50% Asian, which helps to explain why those groups rank 2nd and 3rd.
Costs
Facilities Costs vary enormously from site to site
Let’s move away from the FCI and consider the actual projected facilities cost in dollar terms. High school campuses (891 students per site) are bigger than middle school (635 students per site) and elementary school (340 students per site) campuses and, as we have seen, their higher FCI scores mean they are poorer maintained on average. We would expect them to require more money but the extent of that need is perhaps surprising. Figure 5 is a histogram showing how many schools require how much money.
While most sites require less than $10m over the next five years, nine require more than $30m. Seven of those nine are high schools with the other two being middle schools. These nine sites are projected to require one third of the entire facilities budget, led by Galileo HS ($83 million), Balboa HS ($83 million), Mission HS ($76 million), Lincoln HS ($57 million), and Everett MS ($51 million).
Facilities Cost per student varies widely
For any major system, such as HVAC or power or telecoms, there will be some fixed cost per site and some variable cost that will depend on the number of classrooms or the area of the school. Because of those fixed costs, the average cost per classroom and per student should be lower in a large school than in a small school. That’s the theory. In practice there is enormous variation between schools of similar sizes driven only in part by the different FCI scores of the schools. Figure 6 shows the estimated 5-year facilities cost per student as a function of the number of students in the school.
It is true that expressing the amounts in dollars per student goes some way towards making them seem less exceptional. Galileo HS may require the most in pure dollar terms ($83 million) but, because it has nearly 2000 students, the cost per student is about $44,700 which wouldn’t even make a list of the top 20 most expensive sites. Balboa HS and Mission, because they have fewer students (1259 and 1084 respectively), are much more expensive on a per student basis ($65,800 and $70,600 respectively). The only schools more expensive than Mission on a per student basis are three early education schools and one elementary school (Malcolm X) that is so under-enrolled that it is the size of an early education school.
Elementary schools require less facilities money
Recall that elementary schools are better maintained (i.e. have lower FCI scores) than middle and and high schools. It is thus not surprising that the estimated facilities cost per student over the next five years is much lower for elementary schools than for middle and high schools (see figure 7 below).
The elementary schools have expected costs of $22,000 per student which is 33% less than the $33,000 average estimate for both middle and high schools. Presumably, the estimates for middle and high schools are about the same, even though the high schools are in poorer condition, because the high schools benefit from economies of scale. Meanwhile, the early education schools, which are tiny (only 72 students per site), do indeed have the highest estimated costs (over $40,000 per site).
Why does the district seem to prioritize the elementary schools at the expense of the middle and high schools? At the October board meeting, Dawn Kamalanathan, the head of facilities, said something that explains how it could have happened:
“there has been a rule of thumb to let’s try not to spend more than $30 million per site”
If there is a desire to avoid larger projects, that inevitably means that work on the larger schools is going to get deferred. It feels much better to say “we completed work on six sites” than “we completed work on one site” even if there are more students at the one site than the six. The district needs to change its prioritization approach to give more weight to the number of students than the dollar size of the project.
Under-enrolled schools are expensive to maintain
Let’s narrow the focus to elementary schools. Figure 8 shows the estimated 5-year facilities cost for elementary schools of different sizes.
There are ten schools with 101-200 students and twenty-three schools with fewer than 300 students. They are in about the same current physical condition: the average FCI score for the ten smaller schools is 21% and for the twenty-three larger schools is 22%. The projected 5-year maintenance costs are only 5% apart (in fact, the smaller schools are estimated to require slightly more than the larger ones, $5.9 million versus $5.6 million). That is not particularly surprising because virtually all these schools are significantly under their physical capacities and facilities costs are proportional to the physical capacity of a school, not its current enrollment. But, precisely because they are so under-enrolled, the cost per student of maintaining the smaller schools is much greater ($38,000 per student for the schools with 101-200 students compared to $22,000 per student for the schools with 201-300 students).
Maintain or Close Under-Enrolled Schools?
There was passing acknowledgement at the October board meeting that the Facilities Master Plan, which is being prepared, is linked to the proposed redesign of the elementary school admissions process and to the district’s ongoing operating budget issues. Loosely speaking, the link is as follows:
Declining enrollment means that the district is maintaining more schools and classrooms than it can fill. Poor budgeting means that the district may no longer be able to ignore the problem and may have to close a bunch of schools in order to balance the budget. Given how expensive it is to maintain facilities, it is a waste of money to refurbish sites that are going to be closed. The creation of zones for the new elementary school admissions system is made even harder if there is uncertainty about which schools will remain open. Moving to a new zone-based admissions system will be enormously contentious and closing schools anywhere is always enormously contentious. If both have to be done, it may be better to do them together.
The Facilities Master Plan should therefore give a hint to what the district is thinking in terms of school closures. As an example, consider Malcolm X Elementary in Bayview which requires $17.3 million of work. Buena Vista Horace Mann, Rosa Parks, and Moscone all require even more but they have 596, 384, and 419 students respectively. Malcolm X only has 118 which means the bill works out to nearly $147,000 per student. It would be a gross waste of money to spend that much money to serve so few kids. The only alternatives that make sense are:
close the school building. There is a lot of spare capacity at the neighboring Drew ES, which has only 212 students and, because it was recently refurbished, requires just $0.4 million in facilities costs over the next five years.
find a way to increase the enrollment dramatically so that the cost per student gets down to acceptable levels. Enrollment hasn’t been above 200 since 2004 but it was as high as 415 in 1999-2000. It has fallen so low because few people who live in its attendance area want to send their children there2 and hardly anyone who lives outside the attendance area chooses to send their children there3. In the near term, the only practical ways to increase enrollment significantly would be to close enough other schools that Malcolm X is the only school with space to accommodate those who lose out in the kindergarten admissions lottery or to change the kindergarten admissions process so that people who don’t want to send their children there are forced to.
A similar analysis could and should be done for all schools. It would be easy to save $50 million by closing under-enrolled elementary schools. The district has lost 770 middle schoolers in just the last two years, so Everett’s 605 students could easily be accommodated in other middle schools, thereby saving another $50 million in one shot. The high schools are down 510 students in the last two years. That’s the equivalent of an SF International ($37 million) or an O’Connell ($26 million).
I’m not saying these particular schools should close. My analysis is far too cursory for that. All I’m saying is that the district and the board are being negligent if they don’t look seriously at the opportunity to save money by closing some schools and voters would be right to be skeptical of any proposition that blithely called for maintaining all existing buildings even if they’re half empty.
the number in the presentation was $1.6 billion but the line items only added up to $1.43 billion. I think the discrepancy is because the line-item explanation ignored sites that are not SFUSD school sites such as Administration buildings and charter school sites
According to this board presentation about the new Mission Bay school, 347 TK-5 students lived in Malcolm X’s attendance area (as of Fall 2018). An enrollment of 118 is only 34% of that.
According to SFUSD’s Data Dashboard, 88% of Tk-5 students resided in the 94124 zip code.
Great article. Thanks for pushing on this!