How NUSD's Solar PPA Became a Case Study in Missed Fiscal Opportunity -- and What It Teaches Every Local Government About Bond Fund Strategy
Abstract
This paper examines Novato Unified School District's Power Purchase Agreement with Tesla/SolarCity through two lenses: first, as an investigative account of contractor non-performance and board inaction; second, and more importantly, as an illustration of a systemic failure in how California local governments deploy restricted bond funds. When a district can use authorized bond capital to acquire an asset that eliminates recurring general fund expenses, failure to do so is not merely a missed financial opportunity -- it is a structural choice to keep taxpayers paying more than they need to. The restricted-to-unrestricted conversion framework presented here applies not only to solar energy but to workforce housing, wildfire hardening, and any capital investment where restricted bond funds can generate ongoing operating savings.
1. The Deal
In the mid-2010s, Novato Unified School District signed a Power Purchase Agreement (PPA) with SolarCity (now Tesla) to install solar panels at 10 school sites across the district. The deal was straightforward: Tesla owns and maintains the panels, NUSD buys the electricity at a discounted rate, and Tesla guarantees the systems will hit 100% of projected output over rolling 5-year periods. If they fall short, Tesla owes NUSD money.
When the PPA was signed, the deal made sense on paper. Public entities like school districts could not claim the federal Investment Tax Credit because they have no tax liability. That was the entire rationale for PPAs: a private company could capture the 30% ITC and accelerate depreciation, then pass some of that value to the district through a discounted electricity rate. The district avoided upfront capital costs and got cheaper power.
The tradeoff was significant. The district owned nothing. Tesla captured all ancillary value -- tax credits, depreciation, Renewable Energy Credits, and the asset itself. NUSD was just a customer, buying electricity from a private company operating on public land.
For years, the arrangement saved NUSD hundreds of thousands annually. In the most recent reporting year (August 2023 through July 2024), the 10 sites produced 1.9 million kWh, offset 64% of utility consumption at those schools, and saved the district an estimated $475,300.
2. What Went Wrong
In June 2024, NUSD received a $160,000 true-up bill from PG &E -- more than double the $70,000 bill from the year before. The spike was attributed to new electric HVAC heat pumps drawing more power than the solar panels could offset. CFO Joshua Braff brought the issue to the board, and in September 2024 the trustees approved a $48,000 contract with NV5 Consultants to audit Tesla's performance and register NUSD's Renewable Energy Credits for potential sale.
NV5 presented its findings to the board on May 20, 2025. The report flagged serious problems:
Underperformance
The solar portfolio produced only 83% of its weather-adjusted model -- 17% below the 100% guarantee threshold. Rancho Elementary was the worst performer at 49% below model due to an outage starting April 2024 that was never repaired through the end of the reporting period.
No Maintenance
Tesla provided zero records of any Operations and Maintenance work at any of the 10 sites. The PPA contract (Section VII(b), Exhibit 3) requires Tesla to "operate and perform all routine and emergency repairs to, and maintenance of, the System at its sole cost and expense." This is a direct contractual breach.
Non-Responsive
Tesla stopped responding to NV5's emails entirely. As of May 2025, it had been over four months since Tesla's last reply. NV5 could not verify whether Tesla's performance guarantee true-up calculations were accurate because Tesla would not share the data.
Possible Overbilling
NV5 identified a potential $10,000 billing discrepancy at Rancho Elementary, where Tesla appeared to be charging for electricity the panels were not producing during the outage period.
RECs Ready
On the positive side, NV5 successfully registered all 10 sites with WREGIS and had Q4 2024 and Q1 2025 Renewable Energy Credits ready for sale.
NV5 explicitly recommended a budget amendment so NV5 could continue pursuing Tesla, a meeting to develop a REC sales strategy, and resolution of the O &M records gap, performance guarantee verification, and Rancho billing error.
3. What the Board Did Next
Nothing.
A search of every NUSD board agenda from June 2025 through the most recent meeting on March 24, 2026 -- nearly a full year -- shows zero follow-up items related to Tesla, NV5, or the solar PPA. No contract extension for NV5. No legal demand letter to Tesla. No agenda item addressing the performance guarantee shortfall. No REC sales strategy. NV5's original $48,000 contract expired around November 2025, and no replacement engagement has been brought to the board.
Meanwhile, the PPA contract gives NUSD clear remedies. Section XIII(a)(2) defines a Default Event as "failure of a Party to substantially perform any other material obligation" within 30 days of written notice. The dispute resolution clause (Section XXII(b)) requires mediation demands within 90 days of written notice. The board has not sent formal written notice. None of the cure periods or termination rights have been triggered.
The one firm holding Tesla accountable walked away when its contract expired, and the board let it happen.
4. The Missed Opportunity -- Quantified
What NUSD Left on the Table
Two things changed since 2016 that fundamentally altered the PPA calculus.
First, the Inflation Reduction Act of 2022 created "Direct Pay" (also called "elective pay"), which for the first time allows tax-exempt entities like school districts to claim clean energy tax credits and receive cash directly from the IRS. A school district that owns its solar installation can receive a base credit of 30% of the project cost. With prevailing wage/apprenticeship compliance, domestic content (10% bonus), and energy community location (10% bonus), credits can stack to 50% or higher. The Climate and Community Institute explicitly recommends that "school districts should look beyond the status quo of power purchase agreements."
Second, NUSD already had the capital. Measure G, the $222 million GO bond approved by Novato voters in November 2016, explicitly authorizes energy efficiency improvements.
Model Assumptions
| Parameter | Value |
|---|---|
| System size | 1.3 MW across 10 sites |
| Current annual production | 1.9M kWh (83% of model) |
| PPA start rate (2017) | $0.11/kWh |
| PPA annual escalator | 2% |
| PG &E blended rate (2024) | $0.23/kWh |
| PG &E annual escalator | 4% |
| Year 10 purchase FMV estimate | $2.2M |
| IRA Direct Pay (conservative) | 30% = $660K |
| IRA Direct Pay (optimistic) | 50% = $1.1M |
| Net acquisition cost (30% IRA) | $1.54M |
| Bond financing | 4% over 20 years |
| Annual O &M (district-owned) | $30K |
| Annual REC revenue | $15K (conservative) |
| Panel degradation | 0.5%/year |
Figure 1: Cumulative general fund impact under continued PPA vs. district ownership. Ownership breaks even by approximately 2030 and generates a $6.38 million cumulative advantage by 2036.
Figure 2: Annual cash flow comparison at Year 11 (2027). Under ownership, the bond debt service comes from restricted Measure G funds (hatched), while $452,533 in net savings flows to the unrestricted general fund -- double the $225,219 PPA savings.
Under district ownership financed by Measure G and IRA Direct Pay, NUSD's annual general fund benefit nearly doubles -- from $225,219 under the PPA to $452,533 under ownership. The critical mechanism: bond debt service is paid from restricted capital funds (the Measure G bond fund, supported by property tax assessments), while the full electricity savings and REC revenue flow to the unrestricted general fund. This is the restricted-to-unrestricted conversion in action.
Figure 3: Cumulative missed value since IRA Direct Pay became available in 2023. Total opportunity cost: $1.62 million in foregone savings, uncaptured IRA credits, and lost REC revenue.
5. The Closing Window
The One Big Beautiful Bill Act, signed in 2025, modified the IRA timeline. Under the revised rules:
- Solar projects beginning construction before July 5, 2026 face no placed-in-service deadline and phase out beginning in 2032.
- Solar projects beginning construction after July 4, 2026 must be placed in service by December 31, 2027.
- Starting in 2026, no more than 60% of direct costs can come from prohibited foreign entity materials (primarily Chinese manufacturers), decreasing 5% annually through 2029.
NUSD has approximately 15 months from April 2026 to begin construction and preserve the most favorable credit terms. After July 2026, the window narrows drastically.
The PPA also gives NUSD the right to purchase the solar systems outright at the end of Contract Year 10 -- a window likely approaching in 2026-2027 based on the contract's 2016 execution date. Exercising that option requires 90-180 days advance notice. If NUSD misses the Year 10 window, the next purchase opportunity is at Year 20, roughly 2036-2037.
No board discussion of the purchase option, the IRA Direct Pay opportunity, or the OBBB timeline appears on any agenda.
6. The Bigger Principle -- Restricted-to-Unrestricted Conversion
A GO bond is restricted capital. It can only be spent on authorized purposes. But a strategic board asks one question before every bond expenditure: How do I turn this restricted dollar into an asset that reduces demand on my unrestricted dollars?
NUSD's solar situation is a case study in what happens when that question is not asked. But the principle extends far beyond solar. Any time restricted bond funds can be deployed to acquire an asset that generates ongoing operating savings, failure to do so means the general fund absorbs costs that capital funds could have eliminated.
Figure 4: The Restricted-to-Unrestricted Conversion Framework. Capital bond investments that generate ongoing operating savings effectively convert restricted dollars into general fund relief.
Example 1: Workforce Housing
A school district or county that uses GO bond authority to build or acquire workforce housing for teachers, firefighters, and essential workers converts restricted capital into multiple streams of general fund relief. Lease revenue from the housing units flows to the general fund or a designated housing fund. Reduced recruitment and retention costs lower HR spending -- NUSD and other Marin districts spend significant resources recruiting teachers who cannot afford to live in the communities they serve. Reduced substitute teacher costs follow from better retention. If the housing appreciates in value, the asset itself grows while the bond debt service remains fixed. A shared-equity model adds the benefit of building community stability.
The math: if a district spends $5M in bond funds on workforce housing that generates $200K/year in lease revenue and saves $150K/year in recruitment/retention costs, that is $350K/year in general fund relief from restricted capital -- the same mechanism as solar ownership, applied to a different asset class.
Example 2: Wildfire Hardening
Marin County and its communities face escalating wildfire insurance costs and increasing emergency response expenditures. A county that uses GO bond authority for fire hardening of public facilities -- defensible space, fire-resistant building materials, emergency water storage, vegetation management -- converts restricted capital into reduced insurance premiums (which are general fund operating expenses), avoided emergency response costs, lower FEMA match requirements after events, and protection of assessed property values (which support all property-tax-dependent revenue streams including the bond debt service itself).
The math: if wildfire hardening reduces annual insurance premiums by $500K across county facilities and avoids an average of $300K/year in emergency response costs, that is $800K/year in general fund relief -- while also protecting the property tax base that funds the bond repayment.
In every case, the principle is the same. Restricted capital cannot pay operating expenses directly. But restricted capital can purchase assets that reduce what the general fund must pay. A board that treats bond funds as a separate spending track, disconnected from the operating budget, fails to capture this conversion. A board that asks "how does this bond expenditure reduce general fund pressure?" before every capital allocation decision builds structural fiscal health into every dollar it spends.
For a district like NUSD that has declared fiscal emergency, contemplated school closures, and asked voters for parcel taxes -- Measure A at $251/year, and now a proposed additional $249/year levy on the June 2026 ballot -- the failure to use available bond authority to reduce general fund operating costs is not merely an oversight. It is a structural failure with direct consequences for taxpayers and students.
Figure 5: Total 20-year economic value under PPA vs. district ownership. Under the PPA, Tesla captures $6.8 million in value (tax credits, RECs, asset value, PPA payments). Under district ownership, the district captures $8.7 million, nearly double.
7. Recommendations
1. Re-engage NV5 or equivalent consultant immediately. Resume Tesla accountability, including formal written notice of default under PPA Section XIII(a)(2) to trigger cure periods.
2. Evaluate the Year 10 purchase option. Obtain an independent appraisal of the solar systems. Factor Tesla's documented non-maintenance into the fair market value assessment.
3. Begin IRA Direct Pay planning before the July 2026 construction-start safe harbor. Engage a tax advisor experienced in elective pay for public entities. File IRS pre-registration.
4. Develop a REC monetization strategy. The credits NV5 already registered with WREGIS are idle assets generating zero revenue.
5. Apply the restricted-to-unrestricted conversion framework to all future Measure G expenditure decisions. Every remaining bond dollar should be evaluated for its potential to reduce ongoing general fund costs.
6. Conduct a comprehensive audit of all PPA invoices against actual system production data. Quantify any overbilling beyond the Rancho Elementary discrepancy.
7. Publish all solar performance data, PPA financials, and NV5 findings on the NUSD transparency dashboard. Enable public oversight of contractor performance and district energy costs.
Methodology and Assumptions
The financial model in this paper uses publicly available data supplemented by reasonable estimates where exact contract terms are not available. System size (1.3 MW) and actual production (1.9M kWh) are derived from NV5's May 2025 presentation. PG &E rates are based on the Schedule B-20 commercial tariff applicable to NUSD sites. The 4% annual PG &E escalator reflects the 10-year historical average for California commercial electricity rates.
Certain PPA terms -- specifically the exact starting rate ($0.11/kWh) and annual escalator (2%) -- are estimated from publicly available data on comparable SolarCity school district PPAs of the same vintage and may differ from the executed contract. The PPA template reviewed for this analysis defines these values in Exhibit 4 (Rate Schedule), which was not included in the publicly available version of Contract No. 2016-PPA-000098.
The separate Performance Guarantee document referenced in PPA Section XIV(d) was not available for this analysis. The 100% performance guarantee threshold is based on NV5's presentation and is consistent with standard SolarCity PPA terms.
The Year 10 purchase price is shown as "TBD" in the PPA template (Exhibit 5). The $2.2M fair market value estimate is based on $/W benchmarks for 10-year-old commercial solar systems in California, adjusted downward for the documented maintenance deficiencies. The actual purchase price in the executed contract may differ.
IRA Direct Pay credit percentages (30% base, up to 50% with bonus credits) are based on published Treasury guidance and the Blue Green Alliance analysis of school district eligibility. Panel degradation of 0.5%/year is based on NREL research data cited in Sinovoltaics analysis of crystalline silicon modules.
All projections are presented for analytical purposes and should not be construed as financial advice. NUSD should engage qualified financial, legal, and energy advisors before making acquisition or financing decisions.
Sources
1. SolarCity/Tesla PPA, Contract No. 2016-PPA-000098.
2. NV5 Consultants presentation to NUSD Board of Trustees, May 20, 2025.
3. Climate and Community Institute, "School Solar Ownership Models," December 2024. climateandcommunity.org
4. U.S. Treasury, "FACT SHEET: Inflation Reduction Act Tax Credits Can Fund School District Projects," January 2024. home.treasury.gov
5. Blue Green Alliance, "Federal Government to Reimburse Schools for up to 50% for Clean Energy Projects." bluegreenalliance.org
6. Wipfli, "Schools can still claim clean energy tax credits after OBBB," July 2025. wipfli.com
7. NUSD Measure G Bond Information. nusd.org
8. NRG Clean Power, "How Much Do Solar Panels Cost in California (March 2026)." nrgcleanpower.com
9. NREL solar panel degradation data, cited in Sinovoltaics research.
10. PG &E Schedule B-20 tariff sheet.
Marc Hunter Lewis | Candidate for Marin County Supervisor, District 5 | April 2026