Enrollment Trends

We shouldn't expect a quick recovery to pre-pandemic levels. SFUSD needs to plan accordingly.

Public school enrollment figures for 2022-23 were recently released. We saw a few months ago that San Francisco private school enrollment was down 1.0% in 2022-23 and is down 3.4% since 2019-20. Let’s see how SFUSD and San Francisco’s charter schools did in comparison.

SFUSD

SFUSD enrollment in 2022-23 was down by 419 (0.9%) to 48,785. Since the last pre pandemic year (2019-20), enrollment is down by nearly 4,000 (7.6%).

Individual schools to see big changes in enrollment this year were:

• Among elementary schools, Visitacion Valley was down 43 (14%), and Tenderloin was down 35 (12%). On the other hand, McCoppin (Inner Richmond) was up 27 (14%)

• Everett MS (Castro) was down 94 and nearby Lick MS (Noe Valley) was up 61, probably as a consequence of the reported problems at Everett. Willie Brown (Bayview) was up 61 (31%!) thanks to a huge bump in 9th grade enrollment. Nearby, MLK (Portola) was down 34 (9%) and Visitacion Valley MS was down 21 (6%).

• There were big changes at a number of high schools. Wallenberg HS (Western Addition) was down 87 (14%), Marshall HS (Bayview) was down 69 (13%), Asawa SOTA (Twin Peaks) was down 71 (9%), Burton HS (Portola) was down 84 (7%). On the other hand, SF International (Potrero Hill) was up 108 (37%) and Independence HS (Inner Sunset) was up 68 (50%).

Charter Schools

Aggregate charter school enrollment fell by 5.8% last year. That’s a big decline for one year but the story is very different from school to school.

• New School (North Beach) continues to expand as it adds grades. It’s now up to 439 students in K-7, more than SFUSD’s neighboring Garfield and Yick Wo put together.

• Creative Arts (Western Addition), Mission Prep, and the two Gateway schools (also both Western Addition) are all at or close to their all-time highs.

• Leadership High (Balboa Park) and City Arts and Tech merged but the combined enrollment was down 75 students (14%).

• Edison (Noe Valley) is down 51 students (8%) this year and 142 students (19.4%) since 2019-20. That’s got to be concerning but may not be a crisis because it still has 590 students, more than all but four of SFUSD’s elementary or K-8 schools.

• The four KIPP schools (three in Bayview, one in Western Addition) were each down more than 10% this year. Compared to before the pandemic, the two middle schools are down 121 (38%) and 79 (22%) students and the high school is down 63 (17%).

Cohort Survival Rates

One bit of good news is that cohort survival rates have recovered since the pandemic. Figure 1 shows the cumulative cohort survival rates for the last few years at SFUSD. The grade-to-grade Cohort Survival Rate for a grade is the ratio of the number of students in the grade this year to the number of students in the preceding grade last year. This ratio will be below 100% if the district loses students either to other cities or to private schools and it will be above 100% if the district adds students from those sources. The cumulative cohort survival rate for a grade by multiplying all the grade-to-grade survival rates for earlier grades.

Figure 1: Cumulative cohort survival rates for each grade in SFUSD

It is important to interpret this correctly. The chart shows that grade 12 survival rate for 2022-23 is 94%. That does not mean that this year’s 12th grade enrollment is 94% of the 1st grade enrollment eleven years ago (it’s actually 88% of it). It means that the 12th grade enrollment eleven years from now could be expected to be 94% of this year’s first grade¹ enrollment IF cohort survival rates are unchanged in the meantime.

We can see that, in most years, the survival rates follow a predictable path. The district loses a few students during elementary school, loses a lot at the transition to middle school, and then gains back some or all of the lost students in the high school transition.

The cohort survival rates for 2021-22 stand out as being much lower than other years. It is noticeable that all of the drop in survival rates is concentrated in the elementary school grades. From middle school on, the district lost no more students than in prior years. I have not studied to see if this pattern holds in other districts or is unique to San Francisco. I surmise that 2020-21 survival rates were barely affected by the pandemic because enrollment decisions were made in March 2020, just as the schools were closing, and enrollment was measured in October 2020 when every public and private school was still closed. By the time, enrollment decisions for 2021-22 were being made, all private schools were open and all public schools were closed and showing little inclination to open.

The cohort survival rates for 2022-23 are back in line with the experience of other recent years. That is good news because it means we’re no longer seeing the effects of the pandemic. Except, of course, that the current survival rates have to be applied to a lower enrollment base. A kid who left the city after 2nd grade is going to be missing from the enrollment for the next ten years. In medical terms, the infection has been killed at the cost of some fingers and toes.

Cohort survival rates change over time. An economic downturn could see more people moving out of the city, taking their children with them and driving down cohort survival rates. That’s effectively what happened during the pandemic. Alternatively, a booming economy or increased construction could attract more families with children to the city and send cohort survival rates up. Figure 2 shows the grade-to-grade survival rates for the middle school and high school transition points. Today, only about 86% of fifth graders make it to sixth grade, meaning that one out of every seven fifth graders leaves the district at sixth grade. This is a relatively new phenomenon. In the early 2000s, the 6th grade class size averaged 95% of the preceding year’s 5th grade class size. The private schools see a surge in enrollment at 6th grade and so do the charter schools so this loss demonstrates that SFUSD’s middle school offerings have been losing ground competitively.

Figure 2: How SFUSD’s grade-to-grade cohort survival rates have changed over time

SFUSD always gains students at the 9th grade level but the number that it gains has been going down. In the 2000s, each 9th grade cohort was at least 110% of the preceding year’s 8th grade cohort. It has not reached that level since 2011-12 but it is still greater than 100% every year. Your first guess might be that these extra students are coming from private schools. That was probably true twenty years ago but private school enrollment today is also at its highest in 9th grade.

Market Share

If we think of SFUSD as competing with charter schools and private schools for students, we can examine the market share each has. Figure 3 shows the percentage of students in each grade that attend each type of school. SFUSD has about two thirds of the market in elementary school, dropping to below 60% in middle school, before recovering to just over 60% in high school. Charter schools serve about 4% of elementary school students, 8% of middle school students, and 5% of high school students. Private schools (including parochial) account for about 29% of elementary school students, rising to 34% of middle schoolers and 33% of high schoolers.

Figure 3: Market share, by grade, for SFUSD and San Francisco private and charter schools

Recall from previous posts that private schools having 31% of enrollment doesn’t mean that 31% of San Francisco students attend private school. A sizable number of students from outside the city attend private high schools in the city. (Yes, there are San Francisco kids who attend Crystal Springs and Nueva and schools in Marin but they’re outnumbered by those who come to the city for SI, University, Lick, and specialist smaller schools like Proof, and Sterne.)

Charter schools had much lower enrollment twenty years ago. When SFUSD was gaining students at 9th grade in the early 2000s, it was gaining them from private schools. Today, when SFUSD gains students in 9th grade today, it’s gaining them from the charter school sector.

Enrollment Trends

Figure 4 shows how SFUSD’s enrollment has changed since 2000-01.

Figure 4: SFUSD enrollment compared to 2000-01

Elementary school enrollment went through a steep decade-long decline losing over 5,000 students between 1995-96 and 2006-07. It then gained half of that back before falling again to an all-time low. It’s now 15% below where it was in 2000-01 and 12% below where it was in 2014-15. That’s a loss of over 3,200 students in 8 years.

The decline in middle-school enrollment (24% since 2000-01) is even greater because SFUSD’s middle schools were losing market share during that time. The bounce-back in enrollment that elementary schools saw in the late 2000s and early 2010s became only a temporary bump peaking in 2017-18 when those students reached middle school age.

Considering elementary and middle schools together, SFUSD K-8 enrollment is down 18% since 2000-01. That’s actually a bit better than private schools which are down 20% in that time but the growth of the charter schools means that SFUSD’s market share is down from 67% to 64%, private schools are down from 32% to 30%, while charter schools are up from 1.1% to 5.5%.

The change in market share at the high school level is more dramatic. While SFUSD high school enrollment is down 14% since 2000-01, enrollment in private high schools is up 19% over the same span. A one-year hiccup in 2019-20 is all that was left of the late 2000s/early 2010s elementary school enrollment bounce. SFUSD’s share of the high school market has fallen from 70% to 62% while private schools have risen from 27% to 33% and charters from 2.7% to 5.3%.

Should School Closures Be On The Agenda?

Figure 5 is the same enrollment data rebased to 2005-06 and shown in terms of student numbers rather than percentages. I picked 2005-06 as the reference year because that’s when SFUSD last conducted a round of school closures.

Figure 4: SFUSD Enrollment Change since 2005-06

Total enrollment is down over 5,000 students since then, during which time the district has OPENED three elementary schools (Dianne Feinstein, San Francisco Public Montessori, and Chinese Immersion School at DeAvila, which have 888 students between them), one middle school (Willie Brown; 257 students), and two high schools (Academy at McAteer and SF International with 720 students). That’s not counting June Jordan HS, which opened in 2003 just before the last closures, or the elementary school under construction in Mission Bay.

There’s certainly plenty of capacity to cut, if the district chooses to do so.

Forecasting Future Enrollment

Projecting future enrollment is hard. SFUSD hired professional demographers who produced in 2018 a report forecasting future enrollment needs. The demographers did very detailed work, analyzing each of the planned housing developments in the city, and estimating how many public school students each was likely to produce. Their principal mistake was believing that San Francisco could actually deliver all the promised housing projects on schedule (perhaps they were instructed by their client, SFUSD, to believe those promises). But their estimates of near-term enrollment, which are driven primarily by births and cohort survival rates, were also overly optimistic. As figure 5 shows, the lowest of their scenarios was too high by 1,000 students in the first year and has gotten worse ever since. It’s now too high by nearly 8,000.

Figure 5: Demographers’ Forecasts of San Francisco public school enrollment compared with actual enrollment

Having established that demographic forecasts are really hard, let’s have a go at forecasting SFUSD’s near-term enrollment. My forecast is only going to cover the years for which we already know the births. The model is quite simple and based on only four factors:

1. Transitional Kindergarten growth

2. Births to San Francisco residents

3. The enrollment to birth ratio i.e. the share of births that end up enrolling in SFUSD

4. Cohort survival rates

We’ve already talked about cohort survival rates. Let’s talk about the other factors.

Transitional Kindergarten Growth

Eligibility for transitional kindergarten is being expanded to all four year olds over the next three years. Even when all four-year-olds are eligible, SFUSD’s plan is to have capacity for only 1,420 kids (compared to a kindergarten capacity of around 4,000) because only a minority of eligible kids actually apply. For my purposes, I’ll assume that that all planned TK classrooms are at 100% capacity and that kindergarten enrollment will equal 1st grade enrollment.

Births Are Way Down

Figure 6 shows births in San Francisco to mothers resident in San Francisco. 1990 was the last year in which San Francisco residents gave birth to more than 10,000 kids. There was a steady decline through the 1990s (and, not coincidentally, there was a decline in elementary school enrollment between 1995 and 2005), reaching a low point of 8,119 in 1999. There was a one-time blip in births for the millennium, followed by a modest upward trend (that surely contributed to the increase in elementary school enrollment that we saw from 2007 on) during the 2000s. As recently as 2016, there were 9,061 children born in San Francisco to mothers resident in the city.

Births have fallen precipitously since then and the pandemic did not affect the trend of the decline. In 2021, there were under 7,500 births and last year there were about² 7,100. That represents a decline of more than 21% in just six years. SFUSD has not yet seen the impact of the declining birth rates because the kids born in the last few years are yet to reach school age.

The SFUSD Yield from Births is Way Down

In a closed world, we could forecast kindergarten enrollment by looking at birth rates from six³ years before and 1st grade enrollment from births seven years before. That doesn’t work so well in a city like San Francisco where there is a lot of migration in and out. Many residents who give birth in San Francisco choose to move out before or soon after their children start school and only a portion of them will be replaced by new immigrants to the city. Few people have kids and then decide to move to the city. The exact ratio of SFUSD first grade enrollment to births seven years before fluctuated between 48% and 54% every year from 1990-91 to 2018-19⁴. But, as figure 7 shows, since reaching 54.3% in 2013-14, it has fallen in each of the nine years since and has been just 42% in the last couple of years.

Figure 7: Births in San Francisco to city residents, grouped by the academic year when they would be in first grade, SFUSD’s actual first grade enrollment, and the ratio of enrollment to births. Note: Academic years 2013-14, 2014-15, and 2015-16 contain only 11 months of births because the Kindergarten eligibility date shifted.

Why did the SFUSD yield drop to 42%? Private and charter schools did not see a huge jump in kindergarten enrollment so we have to assume those families left the city.

If births are down 21%, and the fraction of births that later attend SFUSD is down to 42%, does that mean that we should expect first grade enrollment in a few years to be down by 20% or even more? Possibly, but not necessarily. It could be that the births that did occur are disproportionately to people who are more committed to raising their kids in the city. In the five years between 2016 and 2021, the number of children born to residents aged 35-39 barely changed but the number born to those aged 20-24 fell by 33% and the number born to those aged 25-29 fell by 43%!! If older parents are more likely to stay in the city, the fraction who ultimately attend school in the city may rebound. Moreover, although total births were down 17%, births to Asian mothers and White mothers were both down 26%. Since White and Asian parents are more likely to send their kids to private schools, the fraction of kids who attend SFUSD might also increase. And of course historical trends do not have to continue. A recovering economy might attract more families back to the city. The city might actually build some of the housing that it talks about and it might attract families from outside the city, as opposed to families who already live here. If these things do not happen, the size of each cohort entering kindergarten is set to fall for at least the next five years.

Forecasted Enrollment

My forecast only covers the years to 2027-28 because we already know the size of the birth cohorts for those years. Figure 8 shows the enrollment under two scenarios:

• Current Path: Yield from births stays at 42% and cohort survival rates stay as in 2022-23

• Optimistic: Yield from birth rises 2% each year (from 42% in 2022-23 to 52% in 2027-28) and cohort survival rate matches peak year of 2019-20 in all years.

Figure 8: Forecasted enrollment for SFUSD under two scenarios

In the peak year of the optimistic scenario, enrollment reaches 51,000. In the current path, it falls to 46,000. Only time will tell if my scenarios are any more accurate than those of the professionals.

1. I’m using first grade as the base rather than kindergarten because SFUSD’s reported numbers for transitional kindergarten and kindergarten enrollment are unreliable. See the previous post.

2. It’s “about” 7,100 because the official births to residents numbers for 2022 have not yet been published. I estimated it by taking the number of births that occurred in San Francisco in 2022 (which is known) and assuming that the fraction that were born to residents was the same as in prior years.

3. Six, not five. A child needs to be five to enroll in kindergarten. For the 2023/24 academic year, an incoming kindergartner must be five by the end of August 2023. That means they must have been born by the end of August 2018. The kindergarten class will thus consist of students born between September 2017 and August 2018 i.e. six years before the dates of the academic year. The first grade class will consist of kids born between September 2016 and August 2017 i.e. seven years before.

4. In the early 2010s, SFUSD moved the age cutoff for kindergarten from December 2 back to September 2. They made this change gradually. For three successive years, starting with the kindergarten class of 2012-13 the cutoff date was moved back a month, first to November 2, then to October 2 and then to September 2 where it remains. That means the first grade classes of 2013-14 to 2015-16 consisted of those born in an 11-month period instead of the usual 12-month period.