The proposal to close McKenny, Madison, and LP Brown

About this report

This is an independent analysis of the Olympia School District (OSD) proposal to close McKenny, Madison, and LP Brown elementary schools at the end of 2026-27, and of the documents and data used to support it. I am a parent in the district. No organization commissioned or paid for this analysis, and I am not acting on behalf of any group. This effort has been given for the benefit of the community.

The report is written to be accurate first and persuasive second. Figures marked as estimates are mine; all others come from the sources cited. I used an AI assistant (Claude Opus 5.5) to help analyze data and draft text. All analysis has been reviewed and edited by me, and I welcome any corrections. Community members, including members of OSD4All, reviewed drafts and shared comments and documents. Where this report says “we” or “our” (as in “our estimate”), it follows the usual convention of analytical reports; the estimates and conclusions are mine.

About the author

My name is Emerson Christie, and I am the author of this analysis. For anyone wondering about credentials, I hold a doctorate in Environmental and Molecular Toxicology from Oregon State University. More importantly, I am a parent of a 1st grader at Pioneer Elementary School and a kindergartener at Olympia Waldorf School (OWS), which is a private K-8 school with a pre-K program. Our first grader also attended kindergarten at OWS for two years and until recently we had not committed to joining the district. However, trusted family friends that made the move from OWS to Pioneer had informed us that everything was going great and that our kids would likely thrive. That, a great first impression from Pioneer’s principal the year prior, and a mix of personal, philosophical, and financial reasons convinced us that making the move was a good fit for our family. We are, in other words, the kind of family this report argues the district can win back.

Executive summary

Recruitment, not closure, is the answer.

A better option for the board. Our understanding is that the board’s goals are: a balanced budget, a fiscal health score safely above 1.75, and a reserve moving toward 6%. This report shows the board can reach them without closing schools. Closing the forecast gap of about $0.4 million with options from the district’s own plans, and winning back about 50 students a year, brings the reserve to about 6% by 2030-31, roughly one year after closure would at realistic savings (Table 15). That path keeps 741 children in their schools, avoids the risk that closure itself drives families away, and turns a divided community into a partner. It also keeps the decision open: the board adopts enrollment and reserve benchmarks now, the administration reports against them each quarter, and if they are missed, consolidation returns to the table with better data and a record that recruitment was tried.

Currently, OSD proposes closing three elementary schools to save about $3.1 million a year, but the case rests on an unresolved budget question, an incomplete analysis, and a process that may repeat the defect that voided the 2023 closure vote.

The proposal: McKenny students move to Pioneer and Centennial, Madison students to Roosevelt, and LP Brown students to Hansen and Garfield. The board votes November 12, 2026. Hearings are October 12 (McKenny), 15 (LP Brown), and 19 (Madison).

Findings 1-3 test the budget and enrollment case for closure; finding 4 sets out the path forward; findings 5-8 address the 6883 plan's assumptions, the process, and the reserve.

1. The budget is balanced; there is no fiscal emergency. The adopted 2026-27 budget runs a $720,000 surplus with $3.5 million in cuts already taken. The closures would not take effect until 2027-28. The February interfund loan covered the timing of state and levy payments, not a shortfall (see Budget section).

2. The forecast gap is about $0.4 million a year, and alternatives in the district's own plans could cover it. The adopted budget shows spending outpacing revenue by about $0.41 million in 2027-28 and $0.37 million in each of the next two years (Appendix A2). No published forecast supports the roughly $3 million gap staff described on Sept. 24. Options from the district's master plan, efficiency study, and boundary study are estimated at $0.65-0.75 million a year (our estimate); the 6883 report costs none of them.

3. Most of the current decline is a recruitment problem, not a birth-rate decline. In-district births fell 10% between the birth years of the fall 2019 and fall 2026 kindergarten classes; kindergarten fell 31%. The difference is a falling share of local children enrolling, while the district's population grew 17% since 2010. Steeper birth declines will reach kindergarten from fall 2028. The district's forecast assumes the enrollment share never recovers, its own master plan names enrollment recovery as an alternative to closure, and yet the closure materials contain no recruitment study, plan, or cost estimate. Over the full 17 years, elementary enrollment shows no statistically significant decline, and the three closing schools have kept pace with the other eight; the decline arrived across the system after 2020, which points to a district-wide recruitment problem rather than a problem with three buildings.

4. A staged recruitment path meets the district’s goals on nearly the same timeline, without closing schools. Closing the $0.4 million gap with options from the district’s own plans and recruiting 50 more students each year would lift the reserve to about 5.6% by 2029-30 and about 6.3% in 2030-31 (estimate; Table 15). At realistic savings, closure would reach about 6% a year sooner, at the cost of an irreversible decision whose result depends on how many families leave. The students exist: had the district kept its pre-2020 share of local births, this fall’s kindergarten alone would have had 110-155 more students, and the superintendent reported on Oct. 8 that transfers in from other districts rose from 400 to 719 in three years with modest outreach. The target should be set as an enrollment rate (kindergartners per in-district birth) so that it adjusts as births fall, with annual benchmarks and closure held as a fallback if they are missed (see Alternatives).

5. The 6883 plan rests on optimistic assumptions. It holds district enrollment constant, assuming every displaced student stays. On Oct. 8 the district presented its first range, with a low end of about $2.24 million that combines a 5% enrollment decline with two other assumptions. If 5% of the 741 displaced students left and 1-2% of the 1,756 students at receiving schools followed, the district would lose about $0.8-1.1 million a year in revenue (estimate; Table 7). Receiving schools grow 24% to 55% (42% overall) with no added custodial or nurse time. The plan budgets $50,199 a year to maintain three closed buildings, against our estimate of $150,000-260,000, and its transportation savings rest on no route or bus-capacity analysis. Taken together, net savings are more likely $1.3-2.0 million a year than $3.1 million (estimate; Table 8). At realistic class sizes, Centennial and Garfield would be at or over functional capacity (about 92-102%) and Pioneer and Roosevelt would reach about 87-96%, leaving little room for the enrollment swings the district’s own data show. Its figures also do not reconcile: about 18 students are unaccounted for, and per-school savings moved 20-40% from the 2023 estimates.

6. No citizen involvement process has been documented for this proposal (possible legal issue). State law (RCW 28A.335.020) requires a closure policy that provides for citizen involvement before the board considers a closure. Policy 6883 contains no such provision in either its 2011 or its October 2024 version, and the 6883 report documents no community involvement before the proposal reached the board. The 2023-24 efficiency committee cannot fill that gap: its process ended in April 2024, when the board kept Madison and McKenny open after a court voided its December 2023 vote, and LP Brown was not part of that proposal. The 2024 court ruling that voided the last closure vote turned on citizen involvement. By comparison, Bellingham Public Schools, acting under the same statute in the same year, convened a 34-member task force that met eight times in public before any recommendation and published a 154-page analysis for each school proposed for closure (Bellingham Public Schools, 2026).

7. No evaluation of student outcomes or equity impacts has been performed (board commitment). The board's February 2026 budget parameters commit the district to an equity lens, to showing class-size effects, and to evaluating "impacts on student learning" before changing programs. The 6883 report assesses access to programs, not outcomes, and presents demographic tables without analysis. For closures of schools that are not low-performing, research finds short-term learning losses that fall hardest on high-need students and on the receiving schools that absorb the most displaced students. In this plan those are Hansen and Garfield, where low-income students and special education programs would be concentrated (see Who carries the impact).

8. The ending fund balance is already within policy. The adopted 2026-27 budget ends at $8.41 million, 4.4% of spending, inside the 3-7% range Policy 6022 sets. The 6% and 7% targets came from staff; the board's adopted parameter is a fiscal health score, not a reserve percentage. The district has also scored 0 on the state’s days-cash-on-hand indicator every year from 2019-20 through 2023-24, managing cash flow with short-term tools such as the February interfund loan; even a 7% reserve would not have raised that score (Higgins, 2025). Its overall financial health score, 2.6 in 2023-24, was well above the board’s 1.75 parameter. Policy 6022, before being re-written in October 2024, originally reserved the fund balance for emergencies and stated that "recurring problems require permanent solutions." Building the reserve by permanently closing schools departs from that long standing purpose.

What we ask. The board should not approve this plan on Nov. 12, and it has two sound options. First, find the process incomplete and remove the vote from the agenda, because Policy 6883 lacks the citizen-involvement provision state law requires, the same ground that voided the 2023 vote. Second, vote no and direct the administration to return with a complete plan. Either way, no closure should be approved until the district (1) documents citizen involvement for this specific proposal, (2) publishes the structural deficit with and without closures, (3) costs at least three alternatives, (4) adopts, by board resolution, a recruitment plan with a named lead, a budget, and the annual enrollment and reserve benchmarks set out under Alternatives, and (5) completes the student-learning, class-size, and equity analysis the board promised in February. A no vote rests on the board's own authority and needs no legal finding; removing the vote is stronger if the board's attorney agrees the process is defective.

Budget

Two rounds of cuts. This section keeps two sets of cuts separate. The first is already taken: the $3.5 million Reduced Education Plan adopted in June for 2026-27, about 30 positions including the elementary art specialist program, family liaisons, and one more student per class in grades 4-12. The second is proposed: $3.44 million in further cuts that the adopted forecast assumes for 2027-28 without naming them, which the $3.1 million in closure savings would largely supply. Together they come to about $7 million over two years. Unless noted, "cuts" below means the proposed 2027-28 cuts.

The 2026-27 budget is balanced with a $720,000 surplus; the closures are a 2027-28 decision whose necessity depends on a reserve target the district has stated two different ways.

• Scenario A: the closures build the reserve (best supported). The underlying gap is only about $0.4 million a year, and most of the savings push the reserve toward 6-7%. The adopted budget's numbers and the Feb. 26 record both point here.

• Scenario B: the closures balance the budget. Later forecast years already count the closure savings; without them the district faces gaps of about $3-4 million a year, and the closures keep the reserve roughly where it is. Only unpublished figures staff described on Sept. 24 point here.

Scenario A.

The 2026-27 budget runs a $720,000 surplus because the $3.5 million in June cuts, already taken, are built into its spending. For 2027-28, revenue grows $6.1 million (3.1%) while baseline spending grows $7.3 million (3.7%, driven by raises, step increases, and 15% insurance growth), leaving a gap of about $0.41 million (Appendix A2). The 2028-29 and 2029-30 gaps are about $0.37 million each. The spending line grows 3.7%, 2.4%, then 2.2%, closely tracking the inflation index the budget uses (2.7%, 2.4%, 2.2%), which fits a baseline with no cuts subtracted. Had the proposed $3.44 million in 2027-28 cuts already been subtracted from later years, 2028-29 spending would have to jump an out-of-pattern 4.1%. On the district's own numbers, the gap before any cuts is about $0.4 million a year, and the proposed $3.44 million mainly lifts the reserve from 4.4% to 5.8%, toward the budget book's stated 6% goal. (The table also shows smaller cuts in later years than in 2027-28, which does not match how recurring closure savings work; the district should correct or explain it.)

This read on the budget is also supported by records from Feb. 26^th and records and decisions from June. The Feb. 26 record shows the 7% goal was the administration's, not the board's. Staff presented a four-year model labeled "assumes cuts are taken to reach 7% fund balance in 2028-29," requiring about $4.1 million in cuts in 2026-27, rising to about $7.3 million in cumulative cuts by 2028-29 (Appendix A2). Executive Director Davis told the board that "for us to be truly stable and fiscally solvent, we would really be wanting to shoot for closer to 6% or 7%," and described the proposed $4.1 million as about $2 million "to plug the hole" and $2 million to start rebuilding the reserve. Vice President Flores said she had "advocated pretty strongly that we did not adopt an ending fund balance goal"; the board instead adopted the 1.75 fiscal health parameter, with another director noting it does not stop the superintendent from raising the reserve. The February model's no-cuts scenario showed spending exceeding revenue by about $2.1 million in 2026-27, growing to $6.8 million by 2029-30.

On Feb. 26 the board also approved a temporary loan of up to $3.13 million from the Capital Projects Fund (Resolution 683), to be repaid with interest by Aug. 31, because state and levy payments arrive unevenly while payroll does not. Even an 8.3% reserve in 2020-21 left only 28 days of cash, below OSPI's 30-day threshold, so a higher reserve would not fix this. A separate parent analysis of OSPI’s indicators found the same pattern over a longer period: OSD scored 0 on days cash on hand (fewer than 30 days) every year from 2019-20 through 2023-24, as did a third of Washington districts in 2023-24, and a 7% reserve at 2023-24 spending would have raised days cash only from about 17 to about 28.5 (Higgins, 2025). It calls for cash-flow planning, not a larger permanent reserve. These February figures and decisions predate the adopted budget.

By June, higher revenue projections (including the levy increase) and the $3.5 million in 2026-27 cuts adopted in June (already taken) had closed that gap, leaving a forecast gap of about $0.4 million a year. If the district took no action at all, the reserve would drift from 4.4% to about 3.5% by 2029-30, still above the 3% floor. Closing the $0.4 million gap would hold it near 4%; boundary changes (about 496,000 a year, the 2023 study’s estimate) would lift it to about 4.2%, and the uncosted alternatives(0.65-0.75 million a year by our estimates) to about 4.4-4.6% (estimates, as a share of prior-year spending).

Scenario B.

What staff said on Sept. 24 describes a different budget. Finance staff told the board the district had moved away from "any sort of set fund balance of 7%, or 1 month's payroll, we're not doing that," and that the target is a fiscal health score above 1.75. Staff modeled three-year cuts of about $6.8 million to hold a 3% reserve, $9-10 million for about 4.5%, and almost $13 million for 6%; said more than $3 million in further cuts are expected this spring (it is not clear whether this means the closures or cuts on top of them); and said the $3.1 million in closure savings "would put you in that 4% to 4.5% range, keeping us, status quo." Those statements imply an underlying gap near $3 million a year. No published forecast shows that. Either the forecast has deteriorated since June (inflation projections have risen) or staff are using assumptions the board and public have not seen. Whichever it is, the board is being asked to close schools on numbers it has not been given in writing.

If Scenario B is the true budgetary scenario than the board should require the updated forecast behind the Sept. 24 scenarios, published in the same format as the adopted budget, showing the structural deficit with and without closures.

Which budget scenario are we in

The board needs to clarify which budget scenario we are operating under, scenario A or B. The majority of data indicates that Scenario A is the true budgetary scenario for reasons discussed above and summarized in Table 1. It is therefore important to distinguish between the two fiscal components of this scenario; the true funding gap and the reserve. As Table 1 shows, in 2027-28 about $3.0 million of the $3.44 million in proposed 2027-28 cuts add to the reserve rather than closing that year's $0.41 million gap. This is the adopted budget, published in June; no later forecast has been released (calculations: Appendix A2). On Oct. 8, presenting the long-term planning update, the superintendent repeated that “there is no board director directive to me to maintain a 7% fund balance” and that the goal is to stay above 1.75 on the fiscal health indicators. That is consistent with Scenario A. No updated forecast has been published.

Table 1. Adopted four-year forecast, 2026-27 to 2029-30 ($ millions). Revenue, spending before cuts, the resulting gap, proposed cuts, and ending balance. The $3.5 million in 2026-27 cuts adopted in June are already built into 2026-27 spending; the proposed-cuts row shows further cuts for later years, not yet taken. Source: 2026-27 adopted budget book (Appendix A2).

The fund balance policy

Policy 6022 was rewritten on Oct. 24, 2024. The 2011 version reserved the fund balance for one-time emergencies and stated that "recurring problems require permanent solutions." That language defined the reserve's purpose, and the 2024 revision removed it and repurposed the fund as a multi-purpose cash flow tool rather than a true emergency fund. The current version commits to an ending fund balance of 3-7% of the prior year's budgeted spending and requires the superintendent to notify the board outside that range. The adopted 2026-27 budget ends at $8.41 million, 4.4%, within the range. The district's own documents disagree on what the target is (Table 2):

Table 2. Fund balance targets in district documents. The reserve target as stated in each source.

Bottom line

The 2026-27 budget is balanced, with a $720,000 surplus after the $3.5 million in cuts already taken in June. The district’s own adopted forecast shows a modest gap of about 0.4 million a year for the following three years (0.41 million in 2027-28, then $0.37 million and $0.36 million). No published forecast supports the roughly $3 million gap staff described on Sept. 24. Against that gap, about $3.0 million of the $3.44 million in proposed 2027-28 cuts would go to the reserve rather than to balancing the budget. Yet the reserve already meets district policy: at 4.4%, it sits inside the 3-7% range Policy 6022 sets.

The savings themselves are in doubt. The 6883 report assumes every displaced student stays, budgets little to maintain closed buildings, adds no custodial staff at receiving schools, and rests its transportation savings on no route analysis. Together, these assumptions could reduce the promised $3.1 million by roughly a fifth to a third (see Closure). On Oct. 8 the district presented its first range, with a low end of about $2.24 million that combines a 5% enrollment decline with two other assumptions. The best recent evidence suggests the real net savings could be smaller still. A Stanford study of California closures from 2011 to 2019 found that revenue fell about as much as spending, leaving district budgets no better off (Pearman, 2026; see Revenue risks).

OSD’s overall score on OSPI’s financial health indicators was 2.6 of 4 in 2023-24, close to its 13-year average of 2.66 and within one standard deviation of the state average of 3.07 (Higgins, 2025). That is well above the 1.75 the board adopted as its parameter. The current score should be published with the closure decision. Unless the district publishes a forecast showing a larger gap, closing three schools is not needed to balance the budget. It is a choice to grow a reserve that already complies with policy, through savings that may largely fail to arrive.

Enrollment

Elementary enrollment is declining; however, the data do not indicate that closing these three schools is the right response. Table A1.1 lists the data sources used.

Forecasts

The 2023 FLO forecast can be checked against what happened. For 2025-26, two years out, it predicted 3,397 elementary students; the actual was 3,362, a 1% miss. At individual schools it missed by 5.8% on average (Appendix A4), with errors in both directions that cancelled out in the total.

The pattern is expected: students shift between schools through transfers and program moves while staying in the district, so a district total is easier to forecast than how it splits across buildings. However, per building splits and the rate that they change are important to know when considering closures and consolidation. BERK shows the same pattern, already missing individual schools by up to 19% one year out (Appendix A4). Forecasts are reliable about how much elementary enrollment declines and much less reliable about where, and a closure decision is a decision about where. Part of the Hansen miss may reflect FLO and BERK counting students differently.

The actual enrollment shape: rise, peak, decline

K-5 enrollment in the 11 neighborhood schools rose from 3,714 in 2009 to a peak of 4,031 in 2019, then fell to 3,253 in 2026: down 12% from 2009 and 19% from the peak (Appendix A5). A straight line describes this poorly, so the choice of model matters (Table 3):

Table 3. Models of K-5 enrollment, 2009-2026. Fit of competing trend models; a lower AIC means a better fit. Source: our analysis of district September headcounts (Appendix A6).

The best description is a segmented regression: steady growth before 2020 and decline after. A cycle cannot be pinned down with only one peak in 17 years: when the cycle length is left free, the best fit is 15 years, but forcing 20, 25, or 30 years fits only slightly worse and moves the predicted low point anywhere from 2029 to 2036. A generational cycle of about 25 years is plausible and puts the low point near 2032, but the data cannot distinguish it from the alternatives. None of these curves should be used as a forecast (Appendix A6).

The district-wide picture

District-wide enrollment, measured as the annual average of monthly counts (the basis for state funding), peaked at 9,961 in 2019-20 and averaged 9,434 in 2025-26. From 2020-21 to 2025-26 it drifted down about 15 students a year, a trend that is not statistically significant (95% CI -47 to +17; p = 0.26). That is a drift of well under 1% a year. September counts, used in this report only for grade-level and class-tracking analysis, point the same way. Underneath the total, high school is still growing (3,352 in fall 2020 to 3,540 in fall 2026) as larger classes move up, middle school is roughly flat, and elementary is falling (Appendix A7).

The decline will spread upward. Even if kindergarten holds at this year's 513 (including ORLA), middle school falls about 13% by 2033 and high school about 11% by 2034, taking K-12 to roughly 8,400 (down 10%). The district's own projection shows the same direction: K-12 FTE falling from 8,862 in 2025-26 to 8,606 in 2029-30.

School by school, 2009-2026

2019 is the turning point for nearly every school (Table 4). Since 2019, enrollment has fallen significantly across the system (91 students a year, p = 0.002), at the three closing schools combined (29 a year, p = 0.007), and at the other eight schools (62 a year, p = 0.004). Over the full 17 years the picture is different. Elementary enrollment rose and then fell, and its net trend is not statistically significant (p = 0.095). The three closing schools combined show no significant long-term trend (p = 0.36), and their trend is indistinguishable from the other eight schools' (p = 0.78); together they held 23.0% of K-5 enrollment in 2009 and 22.8% in 2026. Individual schools do differ. McKenny is in steady long-term decline (about 7 students a year, p < 0.001), while LP Brown has gained share (8.3% to 9.4%) and Madison has held its share. The steepest long-term losses are at three receiving schools, Hansen, Centennial, and Garfield; Hansen especially needs an explanation for its one-year drop of 20% in 2026 (340 to 271). Only Lincoln and McLane are above their 2019 levels. The pattern matters for the remedy: a decline that arrived across the system at once, after 2020, is a district-wide enrollment problem that recruitment can address, not evidence that three particular buildings have failed. Over 17 years, boundary and program changes affect these numbers (method: Appendix A1 and A9).

Table 4. Elementary enrollment by school, 2009-2026. September K-5 headcount, share of the 11 schools' K-5 total, and linear trends for two periods. Full-period p-values are corrected for year-to-year correlation (Appendix A1). Source: district September headcount by school; our analysis (Appendix A9).

Kindergarten: mainly fewer families enrolling, partly fewer births

• Kindergarten grew about 8 a year from 2009 to 2019, peaked at 684 in 2019, and has fallen about 13 a year since, to 456 in 2026 (Appendix A5).

• Births explain about a third of the drop. In-district births fell from 642 in 2014, the birth year of the fall 2019 kindergarten class, to 576 in 2021, the birth year of the fall 2026 class: down 10% (Washington Department of Health, as reported in the district's master plan). Kindergarten in all programs fell 31% across the same classes (744 to 513). The remainder is a falling share of local children enrolling: the district enrolled about 1.16 kindergartners per in-district birth in 2019 and about 0.89 in 2026. Had the pre-2020 ratio of 1.08-1.16 held, this fall's class would have had about 110-155 more students (Appendix A3).

• Families did not simply delay kindergarten. If they had, 1st grade would have swelled in 2021; growth that year (1.036) was slightly below normal (Appendix A10).

• Olympia's drop is unusually steep. Public kindergarten enrollment nationally fell about 6% from fall 2019 to fall 2024 (Moon, 2025, using NCES data); over the same years, Olympia’s fell 27% (744 to 540, all programs), pointing to local factors.

• Most of these children were born here; where they went is unknown. Some are likely in private school or homeschool, and some families moved away. ORLA does not explain it: counting ORLA, kindergarten still fell from 744 in fall 2019 to 513 in fall 2026, down 31%, while ORLA elementary enrollment held steady. The district holds homeschool declaration counts that would help answer this. Its closure materials contain no plan to recruit or retain these families.

Where the missing students went

ORLA (Montessori, Online, and H-Connect hybrid) students are district students and generate state funding, but they are not in the neighborhood-school files. ORLA elementary has held steady at about 270 FTE since 2023-24, so it does not explain the recent decline. ORLA and hybrid programs absorbed families during COVID, then fell back. Students who are fully homeschooled are not enrolled and generate no funding. Nationally, about half of post-pandemic public enrollment loss reflects population change, about 20% private schools, about 20% homeschooling, and about 10% is unexplained (EdWeek).

Two local data sources would show where Olympia's students went:

• Homeschool declarations. Washington parents file an annual declaration of intent with their district, so OSD holds yearly counts.

• Private school enrollment. OSPI publishes enrollment for approved private schools by school. The district does not track where its residents enroll privately, but area private school counts by year would show whether they grew after 2020.

Families join at every grade: a recruitment and revenue problem

Families keep enrolling long after kindergarten, which means they can be reached, and the students the district is missing represent real money. Kindergarten (including ORLA) averaged about 663 a year from 2014 to 2018; this fall it was 513, about 150 fewer. At about $14,800 per student, each smaller class costs roughly $2.2 million a year, and because each class grows as it moves up, the gap widens to about 175 students by 5th grade. Every year the pattern continues adds another smaller class.

One caution on interpretation: late joining does not explain why Olympia's kindergarten drop (about 27% from 2019 to 2024) is so much steeper than the national 6%. The rate at which students join after kindergarten has not fallen since COVID: no grade transition shows a significant change, and most are slightly higher (kindergarten to 1st: 1.053 before, 1.069 after). The gap is a local enrollment problem, not a timing effect.

• The growth is real, not a timing effect. Tracking the same class, enrollment grows 16-19% between kindergarten and 5th grade, both before and after COVID (individual classes ranged from 14% to 20% outside the COVID years; Appendix A11). These are students joining the district.

• Separately, grades 3-5 still look steady because of timing. They hold the larger classes that entered before the decline; the smaller classes have not reached them yet.

• Gains continue after elementary school. District-wide September headcounts, 2022-2026 (including ORLA), show classes growing 2.2% from 5th to 6th grade, 2.8% from 6th to 7th, 0.9% from 7th to 8th, and 20.7% from 8th to 9th, then easing slightly through 12th (10th to 11th down 4%, 11th to 12th down 6%). The same class is about 24% larger in 8th grade than in kindergarten and about 37% larger in 12th. Only one rate changed significantly after COVID: the ratio of 12th graders to the prior year's 11th graders fell from 0.986 to 0.944 (p = 0.01). With 12 comparisons, one such result can occur by chance, so this is tentative; the district should explain whether Running Start reporting, early graduation, or dropouts account for it. One near-significant change points the other way: the 5th-to-6th grade ratio rose from 0.995 to 1.022 (p = 0.08), about 19 more students a year joining at the start of middle school. Some of these students likely live in the district and attended private, home, or other elementary programs; to that extent they are elementary students the district could have enrolled, the same recruitment gap the kindergarten data show. The 9th-grade jump has averaged 23% a year before COVID and 21% since (Appendix A10), about 130-180 students a year. The district’s demographer, BERK, attributes it to transfers from the Griffin School District, a K-8 district with no high school whose students attend Capital High, but gives no count (master plan, Exhibit 5 note). Griffin enrolls about 580 students across its grades, roughly 60-65 per grade, so even if every Griffin 8th grader enrolled in OSD it would account for well under half of the jump; the district’s budget book also mentions a few high school students from McCleary. The remaining 70-120 students a year need a district explanation: likely sources include students arriving from private middle schools, homeschools, or other districts and students repeating 9th grade.

• The gains demonstrate Kindergarten recruitment is an underlying issue. Families join the district at every stage, which supports a board directive to recruit. The gains are steady but have not grown since COVID, so there is no sign yet of families returning on their own; recovery would take deliberate recruitment.

• Outreach already works. On Oct. 8 the superintendent said signs and social media had produced “a significant increase” in transfers from other districts, from 400 students in 2022-23 to 719 in 2025-26, a net gain of 326, and that the recent decline of about 100 students a year “would be dramatically worse” without them. Gains of that size came without a dedicated recruitment plan, budget, or target.

Population growth has not yet brought young children

The district's population grew 17% from 2010 to 2023 (61,745 to 72,287), and its population of women aged 15-44 grew about 15% from 2010 to 2024. Children under 5 did not follow: after rising through 2017, their number fell about 6% from 2020 to 2024 (state Office of Financial Management estimates, as reported in the master plan). Housing explains part of this: 81% of units permitted in 2020-24 were multifamily, against 17% in 2000-04, and single-family permits fell from about 293 a year to 63.

• The district's forecast assumes the enrollment share never recovers. It projects births by extending the recent decline in fertility rates through 2030, and it applies the post-2020 kindergarten-to-birth ratio (0.96) to every future year. Both are assumptions, not observations. Restoring the pre-2020 ratio would add about 60-100 kindergartners a year to the forecast (Appendix A3).

• Long-term growth potential sits near schools proposed for closure. The district's 2023 consultant identified large tracts of vacant, developable land in the Boston Harbor, LP Brown, McLane, and Hansen attendance areas that "could eventually add new housing and population beyond the 10-year horizon," and called this a "contradiction between 10-year trends and 50-year timeframe potential" (Western Demographics, 2023).

• Regional growth has reversed school closures before. Seattle closed schools in 2007-2009 and later spent $48.1 million to reopen five as enrollment returned (see Finances). Olympia's housing mix differs, so this is a caution against permanent decisions, not a forecast.

What comes next

• Through 2030. A path to recovery is possible with recruitment and decline is also possible without. Those students are already born. However, even if kindergarten stopped falling today, K-5 would drop to about 3,000 by 2030 (95% range 2,860 to 3,164; Appendix A8) as larger older classes move to middle school. Recruitment or an influx of children moving to the district would be required for recovery.

• Birth declines steepen from the fall 2028 kindergarten class. In-district births were 558 in 2022 and 509 in 2023, the birth years of the 2027 and 2028 classes, against 576 for this fall's class. On Feb. 26, 2026, staff cited a preliminary 448 births for 2025; preliminary counts can be revised, and the district's forecast projected 520 for that year. Recovery in the 2030s depends on births stabilizing and/or more families enrolling; the data cannot yet predict when.

• That uncertainty argues against permanent decisions built on a straight-line decline. Closing a building is costly to reverse: reopening typically requires capital funds, often a voter-approved bond or levy (Seattle spent $48.1 million to reopen five schools), and families who leave may not return. The district’s own forecasts do not extend far enough to show whether the space will be needed again.

What the decline does not settle

Declining enrollment calls for a plan to address the decline and for community-centered solutions to the district's financial concerns. It does not establish an immediate need to close schools, which schools to close, how many, or how quickly, or that closure is better than alternatives such as sharing staff, relocating programs, or recruiting families back. Over the full 2009-2026 period, two of the three closing schools are near their 2009 levels. And because the decline so far comes from fewer families enrolling rather than fewer births, part of it may be recoverable, while closures tend to push families away.

Closure

Where do closure savings come from

Category splits in Table 5 are our estimates from the 6883 report's staffing tables and the June 2026 average cost per position; they sum to about $2.9M against the 6883 report's $3.03M school-level subtotal, so treat them as proportions (Appendix A2). The union contract does not prevent teacher reductions; state law permits non-renewal for enrollment decline (RCW 28A.405.210), and the district cut 31 positions for 2026-27.

Table 5. Closure savings by category. Net position changes and estimated annual savings. Source: 6883 report; our grouping (Appendix A2).

How stable the savings estimate is

The totals are similar, but per-school estimates moved 20-40% (Table 6). The 6883 report never mentions or explains the 2023 figures. Our reading is that the changes track staffing assumptions: Madison's $311,000 increase roughly equals the two classroom teachers the 2026 plan cuts, and the McKenny and LP Brown decreases roughly equal the part-time assistant principals the 2026 plan adds at receiving schools, an assumption staff called open on Sept. 24 ("would we otherwise maybe do deans?"). If a single staffing choice moves a school's savings by $200,000-300,000, the estimates should be presented as ranges, not single numbers. On Oct. 8 the district presented its first range, with $3.14 million as the middle case and about $2.24 million as the low end. The middle case still assumes enrollment at the forecast, administrative duties absorbed by existing staff, and only about 20% of utility costs continuing at the closed buildings; in effect it is close to a best case.

Table 6. Per-school savings estimates, 2023 and 2026. How each closure's estimated savings changed between district studies. Sources: 2023 efficiency study; 6883 report.

Leasing or selling the buildings will not help operations: lease money goes to capital or debt funds except amounts covering the leased property's own costs (RCW 28A.335.060, Policy 6020).

Selling would also be hard to undo. Seattle Public Schools closed schools in 2007-2009, saw elementary enrollment rebound within a few years, and spent $48.1 million in levy money to reopen five elementary schools, some closed only a couple of years earlier (Seattle's Child; Crosscut). Seattle still owned those buildings. A district that has sold its sites would need to find new land in established neighborhoods, which is rarely possible at a comparable location, on top of construction costs. A sale also follows its own legal process (notice, hearing, and appraisal under RCW 28A.335.120), separate from the closure vote. The 6883 report says the district currently favors leasing but identifies no prospective tenant, terms, or timeline. Until a tenant is found, the district pays upkeep and security on an empty building, and problems are more likely to go unnoticed (see Facility and operational impacts). The board should require a leasing plan before any vote and commit not to sell while enrollment remains this uncertain.

Revenue risks

Each student brings about $14,800 in state and levy revenue. A Stanford study of California school closures from 2011 to 2019 found that closing districts lost about 288 students on average and that per-student revenue fell by about as much as per-student spending, leaving deficits unchanged (Pearman, 2026). Different funding system, but the 6883 report assumes no enrollment loss. On Oct. 8 the district presented a low-end case of about $2.24 million that combines a 5% enrollment decline with administrative duties filled by certificated staff and continued use of the closed buildings at a higher share of utility costs. It did not say what the 5% applies to or how much of the $0.9 million difference comes from each change. The base matters. Families at the five receiving schools, which already enroll 1,756 students, also face crowding, larger schools, and new boundaries, and some may leave. Table 7 shows how the estimate changes with the base; we use 5% of displaced students plus 1-2% at receiving schools.

Table 7. Enrollment loss under different assumptions. Students lost and annual revenue lost at about $14,800 per student (estimate).

Lost revenue is shown gross. The district’s Sept. 11 example set a $428,000 revenue loss against $360,000 in staff reductions, but it assumed 24 students leaving one school evenly. Here the losses would be spread across five receiving schools and six grades, about one student per grade per school, which rarely allows a staff reduction. The last row is an upper bound, not an estimate.

Table 8 brings the adjustments together. Our estimate and the district’s own low end, once the costs it still omits are added, land in about the same place: roughly $1.3-2.0 million a year, not $3.1 million. Even at $1.3 million the savings exceed the forecast gap of about $0.4 million a year. The point is not that closure saves nothing. It is that the savings are overstated, that most would go to the reserve, and that the final figure depends on how many families leave, a number no one controls.

Table 8. Closure savings after costs the 6883 report omits. Annual, $ millions. Sources: 6883 report; Oct. 8 board meeting; our estimates as labeled.

Not included: LP Brown’s high-poverty learning assistance funding (about $110,000 a year until Hansen qualifies, which is not assured) and one-time costs such as route restructuring, moving, and capital work at receiving schools.

Capacity, utilization, and portables

Capacity is how many students a building can hold; utilization is enrollment as a share of that capacity. Therefore, utilization depends on which capacity is used: the planning capacity proposed by the district or the functional capacity (our term). The district's Long Range Facilities Master Plan (October 2025) elementary capacity assumes 25 students in every general classroom and 10 in self-contained special education rooms and that figure is applied to all grades. However, the capacity that matters is functional capacity: the seats the district can actually fill while keeping K-3 classes small enough to retain its enhanced state K-3 funding. The state funds 17 students per teacher and has required districts to demonstrate that ratio since 2019-20. The teachers’ contract may allow K-3 classes of 25 but filling K-3 rooms to 25 would cost the district part of its enhanced K-3 funding. The district's own budget book gives a 1st-grade class of 19 as an example of meeting it, because non homeroom teachers also count toward the ratio. Every student is funded; however, a K-3 room filled to 25 would break the ratio and cost the district its enhanced K-3 funding, so the 25-per-room capacity cannot actually be used in those grades. In practice, K-3 classes need to average about 19 students, about 6 fewer seats per room. K-3 is four of the six elementary grades, so roughly two-thirds of classrooms lose about 6 seats each; that is what brings the functional capacity in the tables below down to about 22 per room (estimate; Appendix A12).

By the district's own planning definitions, which count 25 per room, elementary schools were 56-63% full in fall 2026. Measured by functional capacity, they were about 72% full (Table 9). Additionally, the district's own studies list more portable seats at the closing schools over time (McKenny, Madison, and LP Brown: 25, 0, and 25 in 2023; 100, 50, and 50 in May 2026), which makes those schools appear less full. Staff should explain why the district carries so many portables, including at schools with unused permanent space, why listed portable capacity at schools proposed for closure has grown, and why portable capacity is counted in planning capacity. Removal is costly, which may explain why portables remain, but portable seats should not count as permanent capacity at schools proposed for closure. Planning capacity also counts only classrooms, not cafeterias, gyms, or restrooms, so it overstates what a campus can serve while operating well.

Table 9. Elementary capacity under each definition. Seats and fall 2026 utilization under the district's, BERK's, and functional definitions. Sources: 2023 efficiency study; master plan (2025); BERK (2026); our estimate.

Receiving schools after closure

Neither state law nor OSPI sets a utilization standard; BERK chose an 80% benchmark. Student counts below are the district's 2027-28 projections; classroom counts and planning capacity are from the district's master plan. As mentioned above, planning capacity assumes 25 students in every room, which the district does not actually staff to in K-3. Functional capacity, mentioned above, is our estimate using about 19 students per K-3 class (the district's own budget book example) and about 27 in grades 4-5, an average of about 22 per room.

On paper there is room; in practice there is not (Table 10). After closure, the receiving schools would enroll 74-89% of their planning capacity, which looks like room to spare. Measured against functional capacity, Centennial and Garfield would be over capacity (enrollment at about 102% of functional capacity), Pioneer and Roosevelt nearly full (about 96%), and only Hansen (about 85%) would have meaningful room, leaving no margin for the enrollment swings the forecasts have already shown. Across all remaining elementary schools there would still be some space overall (enrollment at about 94% of functional capacity, compared with 82% of planning capacity), but most of it would sit at Hansen, not necessarily where students are being sent.

The 22-student figure is an estimate, and the result depends on it. At 24 students per room, Centennial and Garfield would be at about 92% of functional capacity and Pioneer and Roosevelt about 87%; at 21, all four would be at or over capacity. Either way the receiving schools would have little margin.

Other districts define capacity more conservatively. Bellingham Public Schools, in its 2026 closure analysis, counts only 75% of permanent classrooms toward capacity and excludes portables, so that a quarter of rooms stay free as flexible learning space. Even with a steeper decline (about 1,000 fewer elementary students in 2028 than in 2018), it projects enrollment at its receiving schools at 94-97% of that stricter capacity in the first year and 85-88% by 2030-31 (Bellingham Public Schools, 2026, p. 29). OSD’s planning capacity counts every classroom at 25 students and, in some documents, portables. Measured against Bellingham’s definition, Olympia’s receiving schools would be fuller than even our functional figures show.

Table 10. Receiving schools after closure: planning versus functional capacity. Projected 2027-28 enrollment against capacity at 25 and about 22 students per room. Sources: 6883 report projections; master plan classroom counts; our estimate (Appendix A12).

Facility and operational impacts

Classrooms are not the only limit. The district's own master plan warns that "school sites utilizing portables or mini-buildings to meet enrollment needs may have undersized common spaces such as hallways or gymnasiums," and that portables do not add "gyms, cafeterias, restrooms." The 6883 closure report checks classroom capacity but not the lunchrooms, restrooms, recess, health services, and custodial work that grow with every added student.

• Lunch and recess. From 2029-30, state rules require K-5 students to have at least 20 minutes to eat once served (WAC 392-157-125), and schools must provide at least 30 minutes of daily recess, ideally before lunch (OSPI). Policy 6700P commits the district to work toward both. A receiving school near 500 students with an original-size multipurpose room may need more lunch seatings, each fitted after recess and within the day. No schedule is shown.

• Common spaces. Within a cafeteria the fire marshal's posted occupant load caps how many can eat at once. Codes generally do not re-check an existing building when enrollment rises, so "passes inspection" is not the same as "works at 500 students." The district's master plan flags hallways and gyms as potentially undersized at schools expanded with mini-buildings or portables. Hallways matter most because they cannot be enlarged: they carry every student during transitions, dismissal, and building evacuations. Hansen and Pioneer received mini-buildings without whole-school remodels, and Hansen absorbs the most students. The board should ask for room-by-room class assignments for each receiving school in 2027-28 and a fire marshal review of corridor and exit capacity at the higher enrollments. The 6883 report counts $136 million in renovation costs avoided by closing the three schools. It does not estimate what the receiving schools would need, even though it notes an estimated $55 million modernization for Hansen, which absorbs the most students. Cafeteria, gym, restroom, or corridor work needed at receiving schools should be estimated as a one-time capital cost and set against the avoided costs. Capital comes from bonds, not the general fund, so it does not change the operating gap, but it belongs in the same comparison.

• Custodial staffing. Custodial hours stay at 16 a day at every receiving school, even as each adds 6-11 classrooms and 100-190 students (43 newly filled classrooms in total). The closing schools run about 0.7-0.9 custodial hours a day per classroom. Staffing the 43 added rooms at 0.5-0.8 hours each would take about 21-34 hours a day, or 2.7-4.3 custodians, roughly $235,000-375,000 a year at the 6883 report's cost of about $86,700 per custodian (estimate; Appendix A2). That would offset most of the plan's custodial savings (about 5 custodians, roughly $434,000). If the district intends on shifting staff to receiving schools that should be indicated in the plan and costed. If the district staffs custodians by building square footage, which does not change, the added cost would be smaller; the board should ask which method applies.

• Health services. Nurse and health-room staffing stays at 1.0 at each receiving school, including Garfield, which takes on four program classrooms serving preschool and special education students.

• Special education. The developmental learning classroom, life skills, and two preschool programs move. Whether this counts as a "change of placement" for students with IEPs (and so triggers formal written notice) should be confirmed.

• Transportation. About 110 students who now walk to McKenny or Madison would ride buses; average rides lengthen 2-7 minutes. Longer rides can affect breakfast access, which Policy 6700P addresses. The 6883 report (Section 3) projects net transportation savings, not costs: about $116,000 at LP Brown, $17,600 at Madison, $13,200 at McKenny, and $5,000 from McLane students moving to Hansen, after up to $20,000 to restructure four routes. That is plausible only if existing routes have room for every new rider, including the roughly 110 former walkers; the 6883 report gives no route counts or bus capacities and calls the figures preliminary. Its Section 3 per-school figures also add to about $152,000, while its Section 5 summary uses $162,000.

• Free meals. We checked: no student group loses universal free meals; no school crosses the 30% low-income threshold downward.

• High-needs funding. Closing LP Brown gives up its share of the state's high-poverty learning assistance funding, roughly $110,000 a year (estimate; the two qualifying schools received about $206,000 together in 2025-26). The district expects Hansen to qualify in its place in about three years; that outcome is not assured (see Who carries the impact).

• Keeping closed buildings from deteriorating. The district budgets $50,199 a year to maintain all three closed buildings, about $17,000 each. The three schools currently cost $392,953 a year in utilities and maintenance, which the 6883 report says includes insurance (Table A2.3). This figure equals about 13% of current costs. A building kept ready for reuse still needs low-level heat against moisture and freezing, power for fire alarms and security, fire-sprinkler and alarm inspections, roof and gutter upkeep, grounds work, and vandalism response; a realistic upkeep budget is probably $150,000-260,000 (estimate; Table A2.4). The district's report says insurance on closed buildings could be cut by about $30,000 total; that saving is already inside its figure. This lowers net annual savings by roughly $100,000-210,000, about 3-7%. It does not change the overall picture. The larger risk is that an underfunded empty building deteriorates, which erodes the reuse value the district cites as a reason to keep the properties. Major repairs such as roofs are capital costs paid from bond and levy funds, not the general fund. An empty building is at greater risk than an occupied one: leaks, heating failures, and vandalism go unnoticed longer, and many property insurance policies limit coverage for vacant buildings. If the schools close, prompt leasing reduces that risk; leaving them empty on a maintenance budget of about $17,000 a year each maximizes the risk.

6883 Detail and Quality Concerns

Overall, the current 6883 report lacks the necessary detail for the recommendation it makes. A neighboring district shows what the same law can produce. Bellingham Public Schools, also closing elementary schools in 2026 under its own Policy 6883, reached its proposal through a process OSD did not attempt. The comparison is about process, not outcome: Bellingham did recommend closure, but it showed its work (Table 11).

Table 11. Two 2026 closure analyses compared. Sources: Bellingham Public Schools (2026), Analysis: Carl Cozier Elementary School; OSD 6883 closure report (Sept. 17, 2026).

Bellingham’s enrollment decline is deeper than Olympia’s, and its process was far more thorough. If a closure is warranted in Olympia, the board should expect at least this standard before voting.

Closure Impacts

Who carries the impact

All students, families, and school staff primarily carry the impacts of a closure decision. As such, it stands to reason and a requirement is identified in statute, that the district provide for citizen involvement before the board considers a closure. The plan's heaviest effects fall on the district's highest-need students. It raises the number of elementary students attending majority low-income schools from 569 to 906, concentrates special education programs at Garfield, and relies on Hansen crossing a narrow funding threshold. The 6883 report presents the demographic data but does not analyze it.

The standard the board set

The district has no adopted equity policy; its equity policy page set a fall 2023 adoption milestone that has passed. The board's February 2026 budget parameters therefore set the standard. They commit the district to a budget process "through an equity lens," to decisions "grounded in data, including... demographic data," and to resource decisions that "help close opportunity gaps and support students with greatest needs." Program changes are to be made "after evaluating data and surveys and their impacts on student learning." The 6883 report applies none of these tests.

Who is displaced

Shares are 2025-26 figures from the 6883 report's demographic tables; enrollment is the district's September 2026 count.

LP Brown is one of the district's only two schools that qualify for the state's high-poverty learning assistance funding. Madison has the highest shares of English learners and housing-insecure students among the eight affected schools (Table 12). Two closing schools are also more racially diverse than the schools that would receive their students: Madison is 49.7% white and Roosevelt 67.0%; McKenny is 53.5% white and Pioneer 64.6%.

Table 12. Who is displaced. Enrollment and student needs at the three closing schools.

Where need concentrates

Enrollment is the 6883 report's June 2025-26 count before and after consolidation; shares are from its demographic and program tables.

Hansen and Garfield carry the largest load (Table 13). They already have the highest special education shares among the affected schools, they grow by about half, and Garfield adds four program classrooms with no added nurse time. Both would exceed the roughly 300 students that research associates with better results in schools serving mostly disadvantaged students (Leithwood & Jantzi, 2009). In Philadelphia, the largest learning losses after closures occurred at receiving schools that absorbed the most displaced students (Steinberg & MacDonald, 2019).

Table 13. Where need concentrates. Enrollment, low-income, and special education shares at the receiving schools before and after consolidation.

The learning assistance funding

The plan's own revenue recovery depends on concentrating poverty at Hansen.

• What exists today. LP Brown and Garfield qualify for the state's enhanced high-poverty learning assistance funding, which requires a three-year average of at least 50% low-income students. Together they received about $206,000 in 2025-26 (6883 report, p. 31).

• What it can be used for. The money must be spent on added support at the school that generates it (OSPI, 2024b). It cannot close a general budget gap.

• What closure costs. Closing LP Brown ends its share, roughly $110,000 a year (estimate; Appendix A3). The 6883 report states that "additional support for Hansen will be provided in the interim" but does not cost that support, and the savings estimate does not include it.

• Why Hansen's qualification is not assured. The district projects Hansen at 52.7% low-income in one table and 51.7% in another, a margin of 9 to 14 students out of 528. Under the three-year average, Hansen would reach about 46.9% after one year and 49.1-49.8% after two, crossing 50% only in the third year, and only if it stays above 50% every year (Appendix A3). The projection uses 2025-26 data; Hansen enrolled 271 students this fall against 341 in June, a 20% drop of unknown composition. The state formula itself is under review: OSPI's 2027-29 budget request proposes replacing the 50% cutoff.

The state pays this premium because concentrated poverty makes learning harder. The 6883 report treats Hansen's crossing of the threshold as a revenue recovery. An equity analysis would treat the rise in concentrated poverty as a cost to be weighed, mitigated, and funded.

Access Concerns

• Walking to school. About 110 students who now walk would become bus riders (35 at Madison, 75 at McKenny), and average ride times rise by 2 to 7 minutes (6883 report).

• Language access. Policy 4218 classifies closure information as a vital document requiring translation for any language group of 5% or 1,000 people. The district has not confirmed that the 6883 report or hearing notices were translated; Madison is 10.2% English learners.

• Family voice. The board's parameters promise surveys, forums, and an explanation of how feedback was used. The closure materials contain no such record for the affected schools.

What research says about student outcomes

Budget-driven closures are unlikely to improve learning; the research points to a modest, mostly temporary academic cost that falls hardest on disadvantaged students (Table 14).

Table 14. Research on closures, school size, and class size. Main findings, strength of evidence, and relevance to OSD.

The board committed in February to change programs only after evaluating effects on student learning. This research exists and some of it is summarized in Table 14, and the 6883 report does not engage with it.

What a complete impact assessment would include

Each item below is drawn from the board's own February 2026 parameters.

1. A literature review and summary of anticipated impacts to students.

2. Reading and math growth by student group at the closing and receiving schools, with expected effects of consolidation.

3. Projected class sizes by grade at each receiving school.

4. A costed support plan for Hansen and Garfield, including the interim learning assistance support the 6883 report promises.

5. Transportation and walk-zone effects by student group.

6. A record of translation and interpretation for all closure materials.

7. A summary of input from families at the affected schools and how it was used.

Alternatives: Recruitment, not closure, is the answer.

The decline the district is responding to is mostly a falling share of local children enrolling, so the durable fix is to win those families back. The students exist: had the district kept its pre-2020 share of local births, this fall's kindergarten alone would have had 110-155 more students, and families still join at every grade, including about 19 more a year at 6th grade since COVID. Even with falling births, recovering part of that share yields about 50 more kindergartners a year. A staged path would close the $0.4 million gap with options from the district's own plans and add recruitment on top. Alternatives alone would lift the reserve to about 4.5% by 2029-30. Adding 50 more students each year above the district's forecast, which already reflects falling births (150 by 2029-30), counting half of each student's roughly $14,800 in state funding as net of added costs, would lift it to about 5.6% by 2029-30 and above 6% in 2030-31 (estimates; Appendix A3). The board can set annual enrollment and reserve targets and hold closure as a fallback if they are missed, replacing an irreversible decision with measurable conditions.

How closure and recruitment compare

Table 15 projects the ending reserve under each option: the adopted 2026-27 ending balance of $8.41 million, the June forecast gaps, and a 2030-31 gap equal to 2029-30’s. Percentages are of prior-year spending before cuts. Closure savings begin in 2027-28, and the enrollment loss is treated as a one-time step that then holds steady.

Table 15. Projected ending reserve by option. Percent of prior-year spending (estimate; Appendix A2).

Our midpoint uses the middle of each adjustment in Table 7, including 5% of displaced students plus 1.5% at receiving schools. Recruitment rows add the alternatives’ midpoint ($0.70M) and 50 students a year above forecast; the main path counts each recruited student at half of $14,800, conservatively, while lost students count at the full amount (see note to Table 8). Policy 6022 sets a 3-7% range.

What the projection shows. At our midpoint, closure does not lose money. It reaches about 6% in 2029-30, roughly a year ahead of the conservative recruitment path, and about even with it when recruited and lost students are valued alike. Timing is therefore not the argument for closure.

If the district’s figures were right, closure would push the reserve past 7% in 2028-29 and past 9% by 2030-31, well above any target the district has stated. That supports the reading that the closures are chiefly a way to build the reserve (Scenario A).

The risk is one-sided. With midpoint costs, closure savings fall to zero if about 175 students leave, and the reserve starts shrinking once losses pass about 150: roughly 20-24% of displaced students, or 6-7% of the 2,497 students at the eight affected schools. The Stanford study found that California districts that closed schools lost about 288 students on average (Pearman, 2026). That estimate is uncertain (its 95% range runs from about 20 to 560 students), and California differs from Olympia, but if losses here came near that average, the reserve would fall below the 3% floor by 2028-29. The projection may also understate losses: a family that leaves takes younger siblings with it, a student who leaves in elementary school may not return for middle or high school, and closure could deepen the existing decline in the share of local families who enroll.

The two paths fail differently. If recruitment falls short, the reserve stays near 4.5%, within policy, and the board can still consider consolidation later with better information. If closure falls short, the buildings are closed, the families are gone, and the decision cannot readily be reversed. Closure buys perhaps a year on the reserve timeline in exchange for an irreversible decision whose result depends on a number no one controls.

What recruitment can realistically achieve

No rigorous study gives a reliable estimate of how many students a district recruitment effort brings back. The evidence is case reports, several from recruitment vendors, and none fully separates the effect of a campaign from other causes. With that caveat, the evidence suggests 50 students a year is a modest target.

• The district’s own estimate. Transfers in from other districts rose from 400 in 2022-23 to 719 in 2025-26, which the superintendent credited to signs and social media: roughly 80 more students a year from outside the district alone, with no dedicated plan or budget.

• Olympia’s own history. Before 2020 the district enrolled 1.08-1.16 kindergartners per in-district birth; the forecast assumes 0.96. Returning even to the low end of that range would add about 55-104 kindergartners a year (Appendix A3). The target asks the district to recover part of what it had in 2019, not to exceed it.

• The Getting Smart memo estimates the district could attract about 600 additional students in four to five years, about 300 of them in elementary school, and finds about 1,000 K-8 children living in the district who are not enrolled (Getting Smart, 2025).

• St. Paul, Minnesota (about 33,000 students) gained about 400 students in fall 2024, its first increase in a decade, after a year of targeted marketing and program investment at six schools, most of which grew (KSTP, 2024).

• Duluth, Minnesota (about 8,700 students, similar to Olympia) reports more than 200 additional students since starting a targeted digital campaign; the figure comes from the vendor, and attribution is limited (SchoolMint, 2026).

Fifty students a year is about 0.5% of district enrollment: a smaller share than St. Paul or Duluth reported, below what the district’s own pre-2020 enrollment rate implies, and smaller than the gains the superintendent attributes to current outreach. It is not guaranteed, which is why it comes with benchmarks and a fallback.

Benchmarks and implementation

Recruitment succeeds or fails on execution. A fair test needs a plan the board adopts, measures the district does not control, conditions that give recruitment a real chance, and a fallback that a plan never carried out cannot trigger. In place of a closure vote, the board can adopt by resolution:

• A board-adopted plan. By February 2027, the administration presents and the board adopts in open session a recruitment and retention plan with specific actions, dates, a budget line, and a named lead. Strategies could include kindergarten outreach and early registration, a benefits-focused communication campaign, and the program ideas in the Getting Smart memo. Once adopted, the plan is board policy.

• Effort measures, reported separately from results. The board tracks whether the plan is carried out (outreach events held, registration opened on schedule, materials translated, inquiries answered, transfers approved) separately from whether enrollment rises. Results take time; effort can be checked every quarter.

• Enrollment and reserve benchmarks. Kindergarten enrollment of at least 1.00 per in-district birth in fall 2027 and at least 1.05 in fall 2028 (the forecast assumes 0.96); total enrollment at or above the forecast by 50 students in October 2027 and 100 in October 2028; and an ending reserve of at least 4.4% for 2027-28, 4.8% for 2028-29, and 5.3% for 2029-30, with the fiscal health score above 1.75 throughout.

• Independent measures. Enrollment as reported to OSPI each October and births from the Department of Health, plus quarterly leading indicators reported in public session: kindergarten registrations compared with the same point the prior year, transfer requests approved and denied, and the reasons families give when they withdraw, collected through a short exit survey.

• Conditions for a fair test. McKenny, Madison, and LP Brown remain open through at least 2029-30, and boundary changes, transfer caps, program moves, or staffing reductions at those schools require a board vote during the benchmark period. Families do not register children at schools they expect to close, so recruitment cannot be tested while closure remains pending.

• Community oversight. A board-appointed citizen enrollment committee reviews the quarterly reports and advises the board. It also builds the citizen-involvement record that state law would require for any future closure.

• A fallback that tests the plan, not the effort. If the enrollment benchmarks are missed two years running and the board finds the plan was carried out as adopted, the administration returns with a consolidation proposal that documents citizen involvement, costs alternatives, and includes the learning and equity analysis the board promised in February. If the plan was not carried out, the board directs corrective action instead.

The specific numbers are suggestions; the board can set its own. What matters is that the plan is the board’s, the measures are independent, the schools have a fair chance, and a missed target leads to a new decision on full information rather than to an automatic closure.

If births keep falling

Births fall further from 2028, and recruitment slows that decline rather than reversing it. An enrollment-rate target accounts for this. If the district restores its enrollment rate and births still fall, enrollment will decline at a pace the board can see coming, and consolidation can be planned on real data with time for citizen involvement. Meanwhile the district has options short of closure for space it does not need: leasing part of a building to partners such as preschool, child care, or community services; sharing staff and administration across small schools, as in the networks proposed in the Getting Smart memo; and moving programs into underused rooms.

The recruitment target stays within reach as births fall. Returning to the low end of the pre-2020 kindergarten rate (1.08 per birth) yields about 50 more kindergartners a year as long as annual births stay above about 417; the lowest figure cited, a preliminary 448 for 2025, clears it. Births could keep falling beyond the forecast window, since national fertility is at record lows, and every option, closure included, would then face a smaller district. Recruitment ensures that any decline comes from fewer children, not from families choosing to leave.

District’s alternatives

The 6883 report's alternatives section lists only further service cuts (counseling, class sizes, athletics, maintenance, transportation) with no dollar amounts, positions, or student effects. The district's own documents identify other options: using our own cost estimates for a few of these options we have identified at least $0.4-0.75 million a year, enough to cover the $0.4 million structural gap identified from the budget documents, plus enrollment recovery at about $1.5 million a year per 100 students regained. The options in Table 16 come from the district's own master plan, its 2023 efficiency study, the 2026 BERK study, and the board's own questions:

Table 16. Alternatives from the district's own documents. Options the district has raised, with estimated annual effects. Sources: master plan; 2023 efficiency study; BERK (2026); board meetings; estimates are ours where noted.

We deliberately left out options that would disrupt Lincoln, ORLA, or other choice programs; families chose those programs, and the 2023 study's program-consolidation scenarios would be a non-starter.

Most closure savings come from per-building positions. There are approaches some districts use that could reduce those costs without closing buildings, each with trade-offs. They are not formal recommendations for the purposes of this analysis, however, they are worth investigating should the desire to close schools persist despite the rationale and alternatives already listed. For example: centralizing multiple smaller schools administrative activities; one principal serving two small schools, with a head teacher or dean handling day-to-day matters at the other site; and registered nurses covering several schools, with a health-room assistant on site daily. Washington school nurses are registered nurses with a state school-nurse certificate, and OSD's "nurse and health room" staffing may already combine the two roles. The board can ask whether either was considered.

Enrollment recovery is the correct and sustainable path forward. Every option above cuts costs; none brings students back. Since the McCleary school-funding decision and the 2017-18 legislative fix, both state basic education funding and the local levy limit scale with enrollment, so each student is worth about $14,800 a year to OSD. Recovering 100 students would bring in about $1.5 million, nearly half the closure savings. The district should be directed to:

1. Find out where the children are. Report homeschool declarations, ORLA and transition-to-kindergarten enrollment, area private school enrollment, and transfers out by year since 2015.

2. Ask families why. Survey families who left or never enrolled, as the board's February budget parameters already call for surveys before program changes.

3. Publish a recruitment and retention plan with targets, focused on kindergarten entry and on families using homeschool or private school in the early grades.

4. Weigh the closure's effect on that plan. Research links closures to enrollment loss; the 6883 report assumes none.

The pitch to families is straightforward: one of Washington's highest-rated districts, already paid for through their taxes. Niche's 2026 rankings place OSD 14th of 231 Washington districts and first in Thurston County (Niche). Enrollment recovery complements other cost measures rather than replacing them, since births are now falling too.

The district's own master plan points the same way. The October 2025 Long Range Facilities Master Plan, with buildings at about 77% of capacity district-wide, recommends exploring ways to "recapture and grow student enrollment and manage costs," including community schools, community service partners as tenants in schools, and "small school networks," noting that the state's funding model "is designed to support the overhead costs of larger schools." It also suggests expanding ORLA's h-Connect homeschool partnership into a hub-and-spoke model, with other elementary schools hosting homeschool classrooms. It mentions closures only as something that "may need to be considered" alongside boundary adjustments. None of these options appears in the 6883 report.

The comparison that matters (Table 17; the adopted budget points to the first row; staff's unpublished Sept. 24 figures to the second):

Table 17. What the alternatives must cover. How the case for closure depends on the size of the true structural gap.

The only savings unique to closures are utilities and building costs (about $0.4M). Most of the rest of the closure savings come from per-building positions that could, with service trade-offs, be shared without closing a building. The district would argue closure keeps a full-time principal and nurse in each larger school, which is a fair position; it should have been costed against these options.

Policy and legal compliance

Every finding carries one of these labels (Table 18). One item in Table 19 is a possible violation of state law; others are met, gaps in the district’s analysis, or departures from the board's own written commitments rather than law.

Table 18. Evidence labels. How each finding in this report is classified, from possible legal issues to estimates.

Table 19. Policy and legal compliance. Requirements that apply to this closure and whether the record shows them met.

Policy history worth knowing: in October 2024, after the court loss, the board revised Policy 6883 (added item 8, narrowed item 6, added per-school hearings) without adding a citizen-involvement provision. No equity policy has been adopted; the district's equity policy page set a fall 2023 adoption milestone.

Any legal challenge to a Nov. 12 decision must be filed within 30 days (RCW 28A.645.010). Courts generally uphold board decisions unless they are arbitrary and capricious or contrary to law, so procedural defects such as item 1 are stronger grounds than disagreement on the merits. Anyone considering this should consult an attorney before Nov. 12.

Inconsistencies in district documents

None of the items in Table 20 alone defeats the proposal, but together they show figures that were not cross-checked before publication.

Table 20. Inconsistencies in district documents. Figures that differ between district sources.

Background and timeline

This is the district's second attempt in three years; the first was voided in court for lack of citizen involvement, and the policy revised afterward still lacks that provision (Table 21).

Table 21. Timeline of the closure process. Board actions, studies, and court rulings, newest first. Source: district documents and minutes (see References).

References

Research studies and reviews

Beuchert, L., Humlum, M. K., Nielsen, H. S., & Smith, N. (2018). The short-term effects of school consolidation on student achievement: Evidence of disruption? Economics of Education Review, 65, 31-47.

Brummet, Q. (2014). The effect of school closings on student achievement. Journal of Public Economics, 119, 108-124.

Chetty, R., Friedman, J. N., Hilger, N., Saez, E., Schanzenbach, D. W., & Yagan, D. (2011). How does your kindergarten classroom affect your earnings? Evidence from Project STAR. The Quarterly Journal of Economics, 126(4), 1593-1660. https://www.nber.org/papers/w16381

Engberg, J., Gill, B., Zamarro, G., & Zimmer, R. (2012). Closing schools in a shrinking district: Do student outcomes depend on which schools are closed? Journal of Urban Economics, 71(2), 189-203.

Filges, T., Sonne-Schmidt, C. S., & Nielsen, B. C. V. (2018). Small class sizes for improving student achievement in primary and secondary schools: A systematic review. Campbell Systematic Reviews, 14(1), 1-107. https://pmc.ncbi.nlm.nih.gov/articles/PMC8428040/

Johnston, W. R., Engberg, J., Opper, I. M., Sontag-Padilla, L., & Xenakis, L. (2020). Illustrating the promise of community schools: An assessment of the impact of the New York City Community Schools Initiative (RR-3245-NYCCS). RAND Corporation. https://www.rand.org/pubs/research_reports/RR3245.html

Krueger, A. B., & Whitmore, D. M. (2001). The effect of attending a small class in the early grades on college-test taking and middle school test results: Evidence from Project STAR. The Economic Journal, 111(468), 1-28.

Leithwood, K., & Jantzi, D. (2009). A review of empirical evidence about school size effects: A policy perspective. Review of Educational Research, 79(1), 464-490.

Moon, T. (2025, August 5; updated 2026, July 2). K-12 public school enrollment declines explained. FutureEd. https://www.future-ed.org/k-12-public-school-enrollment-declines-explained/

National Education Policy Center. (n.d.). [Review of the Campbell systematic review of class size research]. https://www.nepc.colorado.edu/node/13020

Pearman, F. A., II. (2026, May). The fiscal consequences of school closures in California: Evidence from a statewide synthetic difference-in-differences design. Getting Down to Facts, Stanford University. https://www.gettingdowntofacts.com/sites/default/files/The%20Fiscal%20Consequences%20of%20School%20Closures%20in%20CA%20Evidence%20from%20a%20Statewide%20Synthetic%20Difference-in-Differences%20Design.pdf

Steinberg, M. P., & MacDonald, J. M. (2019). The effects of closing urban schools on students' academic and behavioral outcomes: Evidence from Philadelphia. Economics of Education Review, 69, 25-60.

The Policy Scientist. (n.d.). How do school closures affect students' educational and labor market outcomes? https://www.thepolicyscientist.com/post/how-do-school-closures-affect-students-educational-and-labor-market-outcomes

Thomas B. Fordham Institute. (n.d.). The impact of school closures: A deep dive into the research. https://fordhaminstitute.org/national/commentary/impact-school-closures-deep-dive-research

District documents and commissioned studies

BERK Consulting. (2025). Olympia School District long-term enrollment forecast [Appendix to the Long Range Facilities Master Plan].

BERK Consulting. (2026, May 22). Olympia School District attendance area boundary study 2026.

Integrus Architecture. (2025, October). Olympia School District long range facilities master plan 2025-2035.

Olympia School District. (2008). 2007-2010 budget balance sheet and 2008-2012 budget outlook.

Olympia School District. (2015, July 14). Superintendent's proposal: Final public hearing [2015-16 budget].

Olympia School District. (2020-2024). F-195 budget filings, 2020-21 through 2024-25.

Olympia School District. (2024). 2024-25 budget book.

Olympia School District. (2025). 2025-26 budget book.

Olympia School District. (2026a, February 26). Board budget process and parameters.

Olympia School District. (2026b, February 26). OSPI financial health indicators model through SY 2025, with four-year projections.

Olympia School District. (2026c, February 26). Cover memo and Resolution No. 683: Interfund loan from Capital Projects Fund to General Fund.

Olympia School District. (2026d, June 18). Board budget first reading, 2026-27.

Olympia School District. (2026e). 2026-27 budget book.

Olympia School District. (2026f, September 17). Policy 6883 closure report.

Olympia School District. (2026g). Board of Directors meeting minutes and transcripts [Feb. 26, Aug. 6, Aug. 13, Aug. 27, Sept. 10, Sept. 24, and Oct. 8 2026 ].

Olympia School District. (2026h). Enrollment data, 2009-10 to 2026-27 [Data set]. OSD enrollment dashboard.

Olympia School District Board of Directors. (various dates). Board policies 1260, 4218, 4220 (2011 and 2024), 5000, 6020, 6022 (2011 and 2024), 6114, 6700P, 6883 (2011 and 2024), and 6922. BoardDocs.

Western Demographics. (2023, November 21). Olympia School District school facility efficiency report: CAC findings.

Laws and regulations

Office of Superintendent of Public Instruction. (2024a). Bulletin No. 025-24: Recess requirements. https://ospi.k12.wa.us/sites/default/files/2024-04/bulletin_025-24.pdf

Office of Superintendent of Public Instruction. (2024b). Learning Assistance Program (LAP) guide. https://ospi.k12.wa.us/sites/default/files/2024-08/lap-guide.pdf

Office of Superintendent of Public Instruction. (2026). 2027-29 biennial operating budget request: Learning Assistance Program. https://ospi.k12.wa.us/sites/default/files/2026-09/2027-29-biennial-operating-budget-request_lap_lea.pdf

Wash. Admin. Code § 392-157-125 (2025). https://lawfilesext.leg.wa.gov/law/wsrpdf/2025/10/25-10-005.pdf

Wash. Rev. Code § 28A.335.020 (school closures). http://app.leg.wa.gov/RCW/default.aspx?cite=28A.335.020

Wash. Rev. Code § 28A.335.060 (rental and lease proceeds). http://app.leg.wa.gov/RCW/default.aspx?cite=28A.335.060

Wash. Rev. Code § 28A.335.120 (sale of real property).

Wash. Rev. Code § 28A.405.210 (certificated employee contracts and nonrenewal).

Wash. Rev. Code § 28A.645.010 (appeals from board decisions). https://codes.findlaw.com/wa/title-28a-common-school-provisions/wa-rev-code-28a-645-010/

News, data, and other sources

Bellingham Public Schools. (2026, October 1, revised). Analysis: Carl Cozier Elementary School. Includes the Facilities Planning Task Force 2026 recommendation (May 27, 2026).

Education Week. (2024, June). What's going on with public school enrollment? All the big questions, answered. https://www.edweek.org/leadership/whats-going-on-with-public-school-enrollment-all-the-big-questions-answered/2024/06

Education Week. (2024, August). Public schools launch marketing campaigns to compete with school choice. https://www.edweek.org/leadership/public-schools-launch-marketing-campaigns-to-compete-with-school-choice/2024/08

Getting Smart. (2025, October). Olympia School District: Recommendations for enrollment growth and small school sustainability. Memo commissioned by OSD4All.

Higgins, B. (2025, June). Olympia School District ending fund balance & OSPI school district financial health indicators report. Unpublished analysis.

KING 5. (2024, March). [Judge rules Olympia School District's closure vote invalid]. https://king5.com/article/news/local/olympia/olympia-school-district-vote-invalidated-school-closure/281-d04d6aef-2d26-47bf-ae9a-fa0b1a282c22

KSTP. (2024, November 1). SPPS sees increase in enrollment for first time in a decade. https://kstp.com/kstp-news/top-news/spps-sees-increase-in-enrollment-for-first-time-in-a-decade/

Lilly, D. (2009, November). Time to consider reopening Seattle high schools. Crosscut. https://crosscut.com/2009/11/time-consider-reopening-seattle-high-schools

Niche. (2026). Olympia School District. https://www.niche.com/k12/d/olympia-school-district-wa/

Niche. (2026) Griffin School District profile. https://www.niche.com/k12/griffin-school-olympia-wa/

SchoolMint. (2026, September 2). How Duluth Public Schools increased enrollment by 200 students with targeted digital advertising (vendor case study). https://schoolmint.com/how-duluth-public-schools-increased-enrollment-by-200-students-with-targeted-digital-advertising/

Seattle's Child. (n.d.). [Seattle Public Schools reopens closed elementary schools]. https://www.seattleschild.com/?p=1210

Vinson, T. (2026, March 3). Olympia parents worry school closures are imminent as district plans $4M in cuts. The Olympian.

Washington State Department of Health. (2026). 2024 vital statistics highlights. https://doh.wa.gov/sites/default/files/2026-04/422-099-2024-VitalStatHighlights.pdf

Washington State Health Care Authority. (n.d.). Thurston County characteristics: Births. https://www.hca.wa.gov:443/assets/program/thurston-county-characteristics-births.pdf

Appendix A: Statistical methods and results

This appendix reports every statistical calculation used in the body of the report. Raw enrollment data, the statistical workbook, and the code and checks that reproduce every figure are available at https://github.com/EmersonCodeProjects/olympia-closures/

A1. Data and methods

Data. District enrollment files copied from the OSD enrollment dashboard: (1) district-wide monthly headcount, all grades, 2009-10 to 2026-27; (2) monthly headcount by grade for the 11 neighborhood elementary schools; (3) September headcount by grade and school, 2009-2026; (4) ORLA K-12, middle school (grades 6-8), and high school (grades 9-12) monthly headcount by grade. Financial figures come from the adopted budget books, the June 18, 2026 budget first reading, the Feb. 26, 2026 financial health model, and the 6883 report. Capacity figures come from the Long Range Facilities Master Plan (2025), BERK (2026), and the 2023 efficiency study.

Data cleaning. Stray entries listing 0-3 students in the wrong grade band were omitted. School-level and grade-level analyses use September headcount (the district's staffing basis); district-wide totals use the annual average of monthly counts (the state funding basis). The 2026-27 school year has only a September count and is treated as provisional.

Methods. Trends are ordinary least-squares regressions of headcount on year, reported with 95% confidence intervals and two-sided p-values. Full-period trends (2009-2026) use Newey-West standard errors, which correct for year-to-year correlation in a series that rose and then fell; without the correction, p-values for those trends would overstate certainty. Shorter segments (2019 or later) use ordinary standard errors, because they are too short for the correction to be reliable. Results are called significant at p < 0.05 throughout, with no correction for multiple comparisons. Period comparisons of grade-advancement ratios use Welch two-sample t-tests. Projections use the cohort-survival method: each future grade equals the prior year's grade below multiplied by an advancement ratio. Uncertainty is estimated by bootstrap: 20,000 draws, each resampling a whole year's set of advancement ratios and adding 4% random variation to kindergarten. Model fit is compared with R² and Akaike's Information Criterion (AIC). With six post-COVID years, statistical power is limited; non-significant results should be read as inconclusive, not as evidence of no change. Significant results are stronger evidence but not proof: they mean the pattern would be unlikely if there were no real trend, yet with few years of data their size is imprecise (see the confidence intervals), and where many comparisons are made, an isolated significant result may still be chance.

Table A1.1. Enrollment data sources. What each source counts and how it should be read.

A2. Financial calculations

Table A2.1. Forecast gap, June 2026 adopted forecast ($ millions). Revenue and spending growth behind the structural gap.

Table A2.2. Feb. 26, 2026 financial health model ($ millions). The staff model built to reach a 7% reserve, with and without cuts.

Closure savings by category (our estimate). Staffing changes from the 6883 report multiplied by June 2026 average costs per FTE: administrators $224,124; certificated staff $150,030; custodians $86,744; office staff $81,440; paraeducators and food service $52,522. Classified hours are converted at 8 hours per FTE. Results: certificated about $1.48M, classified about $1.0M, and utilities and building budgets $403,371 (from the 6883 report), about $2.9M in total against the 6883 report's $3.03M school-level subtotal.

Custodial staffing at receiving schools. 43 newly filled classrooms (Centennial 7, Garfield 10, Hansen 11, Pioneer 6, Roosevelt 9) × 0.5-0.8 hours per room per day = 21-34 hours, or 2.7-4.3 FTE × $86,744 ≈ $235,000-375,000 a year.

Closed-building upkeep. Current utilities and maintenance of $392,953 × 30-50% = $118,000-196,000, plus 0.25-0.5 custodial FTE ($22,000-43,000) and security ($10,000-20,000), for about $150,000-260,000 a year, against the district's $50,199.

Table A2.3. Current utilities and maintenance at the closing schools. Annual costs, including insurance. Source: 6883 report.

Table A2.4. Estimated annual cost to keep a closed building ready for reuse. Our assumptions per building (estimate; Appendix A2).

Revenue per student. About $14,800 in state and levy revenue per student (June 2026 budget first reading); 100 students ≈ $1.48M a year.

A3. Equity, enrollment-risk, and payroll calculations

Attrition sensitivity (estimate). The 6883 report holds district enrollment constant. Displaced students total 741 (September 2026: McKenny 275, LP Brown 306, Madison 160). Lost revenue uses about $14,800 in state funding per student and is gross, before any staffing reductions that fewer students would allow.

Table A3.1. Attrition sensitivity (estimate). State funding lost if some displaced students leave the district.

Receiving-school growth. From the 6883 report's June 2025-26 counts before and after consolidation (pp. 3-5): Centennial 415 to 574 (+38.3%), Pioneer 395 to 490 (+24.1%), Roosevelt 349 to 521 (+49.3%), Garfield 256 to 378 (+47.7%), Hansen 341 to 528 (+54.8%). Combined: 1,756 to 2,491 (+41.9%).

Students in majority low-income schools. Before: LP Brown 313 + Garfield 256 = 569. After: Hansen 528 + Garfield 378 = 906. Both on the 6883 report's 2025-26 basis.

Learning assistance share (estimate). The two qualifying schools received about $206,000 in 2025-26. Allocating by enrollment, LP Brown's share is $206,000 x 313 / (313 + 256) = about $113,000. The state formula uses each school's prior-year average enrollment, so the actual figure may differ.

Hansen threshold. At 528 students, 50% is 264 low-income students. The district's 52.7% projection is 278 (14 above the line); its 51.7% projection is 273 (9 above). The three-year average assumes 2025-26 and 2026-27 remain at 44.0%, which Hansen's 20% enrollment drop this fall makes uncertain.

Table A3.2. Hansen's three-year low-income average after consolidation. Projected average under the district's two low-income estimates.

One month's payroll. Salaries and benefits are 83% of general fund spending (2026-27 budget). One month is $195.67 million x 0.83 / 12 = $13.5 million: 6.9% of 2026-27 budgeted spending, or 7.1% of 2025-26 budgeted spending ($190.15 million), the basis Policy 6022 uses.

Births and the kindergarten-to-birth ratio. In-district births (Washington Department of Health, master plan Exhibit 10): 2014, 642; 2019, 585; 2020, 557; 2021, 576; 2022, 558; 2023, 509; 2024, 523. Each birth year enters kindergarten five years later. Births for the 2019 and 2026 classes: 642 to 576 (-10.3%). Kindergarten, all programs: 744 to 513 (-31.0%). Ratio: 744 / 642 = 1.16 in 2019 (matching the master plan's 1.16 for 2019-20); 513 / 576 = 0.89 in 2026. At the pre-2020 ratios of 1.08-1.16, 576 births yield 622-668 kindergartners, 109-155 more than enrolled. Applied to the forecast's projected births of 459-520 a year (2025-2030), the gap between those ratios and the forecast's 0.96 is about 55-104 kindergartners a year.

Staged recruitment path (estimate). Starting from the adopted 2026-27 ending balance of $8.41 million, each year adds the alternatives' midpoint ($0.70 million) less the forecast gap ($0.41, $0.37, and $0.36 million), plus recruited students at $7,400 net each (half of about $14,800 in state funding, assuming most join existing classrooms). Recruitment adds 50 students a year: 50, 100, and 150 above forecast in 2027-28 through 2029-30. Ending balances: $9.07 million (4.6% of prior-year spending), $10.14 million (5.0%), and $11.59 million (5.6%). Alternatives alone reach $9.37 million (4.5%) in 2029-30. Assuming a 2030-31 gap like 2029-30's and 200 recruited students, 2030-31 reaches about $13.4 million (6.3%).

Reserve projection by option (estimate). Each path starts from the adopted 2026-27 ending balance of $8.41 million and subtracts the June forecast gaps ($0.41, $0.37, and $0.36 million, with 2030-31 assumed equal to 2029-30). Closure paths add net annual savings from 2027-28: $3.14 million (6883 report), $2.24 million (district low end, Oct. 8), $1.66 million (our midpoint: 3.14 less upkeep $0.155, custodial $0.305, transportation $0.08, and 63.5 students at $14,800), and $1.32 million (our higher-cost case). Percentages divide by prior-year spending before cuts ($195.67, $202.93, $207.89, and 212.52 million). With midpoint costs ($0.54 million), savings reach zero at about 175 lost students and fall below the forecast gap at about 150. The stress test applies Pearman’s average loss of 287.67 students (standard error 137.27).

A4. Forecast accuracy

Table A4.1. 2023 FLO forecast for 2025-26 compared with actual 2025-26 average headcount. School-level forecast errors two years ahead.

Mean absolute school-level error 5.8%; standard deviation of errors 8.6%. Part of the Hansen error may reflect differences in how FLO and BERK count students.

Table A4.2. BERK 2026-27 forecast compared with September 2026 headcount. School-level forecast errors one year ahead.

BERK forecasts annual averages while these counts are September headcounts; district-wide, the two differ by about 5 students in a typical year and by up to about 70, so a small part of the gap is definitional.

A5. September K-5 headcount by grade, 11 neighborhood elementary schools

Table A5.1. September K-5 headcount by grade, 11 neighborhood schools, 2009-2026. Source: district enrollment data.

Table A5.2. K-5 enrollment trends by series and period. Linear trends with 95% confidence intervals.

A6. Model comparison, K-5 total, 2009-2026

The cycle models fix the period at the stated length (four fitted parameters); when the period is left free, the best fit is 15.1 years (AIC 162.1). With one observed peak, the cycle length is not identifiable.

Table A6.1. Model comparison, K-5 total, 2009-2026. Fit statistics for trend and cycle models.

A7. District-wide enrollment, all grades

Table A7.1. District-wide enrollment, all grades, 2009-10 to 2026-27. Annual average and September headcount. Source: district monthly enrollment data.

Table A7.2. District-wide enrollment trends. Linear trends in annual average headcount by period, with 95% confidence intervals.

A8. Projections

Table A8.1. K-5 projections, 11 neighborhood schools (bootstrap cohort-survival, 20,000 draws). Each draw resamples a whole year's set of advancement ratios (2010-2019 and 2022-2026) and adds 4% random variation to kindergarten. Values are medians with 95% intervals.

Table A8.2. K-12 projections, all programs including ORLA. Deterministic projection using 2022-2026 mean advancement ratios.

A9. Per-school trends

School-level trends (2009-2026 and 2019-2026, with p-values) appear in the table under "School by school, 2009-2026" in the enrollment section, using September K-5 headcount by school and the method in A1.

A10. Grade-to-grade advancement ratios

Ratio = students in grade g+1 this fall ÷ students in grade g the prior fall. Welch two-sample t-tests compare periods; the 2020 and 2021 transitions (COVID) are excluded. Advancement ratios are year-over-year growth rates and show little serial correlation (average lag-1 autocorrelation −0.20 for K-5 and −0.01 for K-12), so no correction is applied; adjusting for effective sample size leaves every K-5 result unchanged. With 10 and 5 years per period, these tests detect only large changes, so a non-significant result indicates no detectable change rather than equality.

Table A10.1. Grade-to-grade advancement ratios, K-5, 11 neighborhood schools. Mean ratios before and after COVID, compared with Welch t-tests.

The 2021 kindergarten-to-1st ratio, a direct test of delayed kindergarten entry, was 1.036, slightly below the pre-COVID mean.

Table A10.2. Grade-to-grade advancement ratios, K-12, all district programs including ORLA, 2009-2026. Mean ratios before and after COVID, compared with Welch t-tests.

A11. Same class tracked from kindergarten to 5th grade (11 neighborhood schools)

Table A11.1. Kindergarten classes tracked to 5th grade, 11 neighborhood schools. Growth of each class between kindergarten and 5th grade.

Classes entering in 2016, 2018, and 2019 passed through the 2020-21 COVID transitions, when classes shrank instead of growing.

A12. Capacity calculations

Planning capacity (district method, master plan): 25 students per general classroom and 10 per self-contained special education room. Functional capacity (our estimate): K-3 classes average about 19 students (the district's own example of meeting the state's 17:1 K-3 funding ratio) and grades 4-5 about 27, so a K-5 room averages (4 × 19 + 2 × 27) ÷ 6 ≈ 21.7 students.

Functional capacity = general classrooms × 21.7 + special education rooms × 10. Example, Centennial: 22 × 21.7 + 2 × 10 ≈ 497. District-wide (11 schools), 204 general and 12 special education rooms give about 4,550 seats today. After closing LP Brown (18 rooms, 1 special education), Madison (12), and McKenny (20), about 3,470 seats remain; projected 2027-28 enrollment of about 3,270 equals about 94% of functional capacity and about 82% of planning capacity (3,970 seats).