The Bessent/Warsh Accord: Part III

This Time the Coming Late 2028 to March 2029 50% Decline Is the Buying Opportunity of the Next Decade.

Preamble

This is the third installment of the Accord series. "The Bessent/Warsh Accord" codicil of August 10, 2026 described the machine: a re-integration of monetary and fiscal power whose output is the total dominance of the US dollar. "Genius Move," published August 20, 2026, described the play: the sovereign's marginal borrowing cost moved off the rate the bond market sets and onto the rate the Accord administers, funded by a statutory bid drawn from roughly $45 trillion of trapped foreign savings.

This report describes the consequence that matters most to an investor with a decade horizon. The break this framework projects for late 2028 through roughly March 2029, a decline on the order of 50 percent from the peak, will be a floor rather than a trapdoor. It will be the buying opportunity of the next decade, precisely because the mechanism that turned 1929's crash into 1932's abyss has been unbolted from the system.

That claim demands care, because the most expensive words in market history are "this time is different." So this report is built backwards from the failure case. Part I reconstructs what happened to the investor who bought the dip in 1930, the single most costly buy decision of the twentieth century. Part II identifies exactly why that purchase failed: not the crash, but the brake applied after it. Parts III through V then argue the affirmative case, that the decline still comes, the floor beneath it is structural, and the arithmetic of the entry is generational. Part VI closes with the checklist that separates the buy from the trap in real time, because a framework that cannot tell you when it is wrong is not a framework. It is a faith.

Everything that follows is analog reasoning, stated as our framework's expectation. Dates and magnitudes are illustrative conventions of the 99-year clock, not forecasts.

Part I: The Ghost at the Table. What Buying the Dip in 1930 Did to You

Begin with the investor this report must answer: the intelligent, disciplined buyer of April 1930.

He had every reason to feel vindicated. The crash of September through November 1929 had taken the Dow from its September 3 peak near 381 down 48 percent to 198.60 on November 13. Then the market did what markets are supposed to do after panics. It recovered. From the November low the Dow rallied 48 percent over five months, reaching a secondary closing peak of 294.07 on April 17, 1930, a level only about 20 percent below the all-time high. The economy had absorbed a financial shock, the Fed had cut rates, and John D. Rockefeller was buying. The 1930 dip-buyer was applying the exact lesson every prior panic had taught, in 1907 and 1914 and 1921: buy the break, hold through the noise.

He was then destroyed more completely than any equity investor in American history. From the April 1930 rally high, the Dow fell almost without interruption for twenty-seven months, closing at 41.22 on July 8, 1932, down 89 percent from the 1929 peak and roughly 86 percent from the point where the dip-buyer had committed his capital. The index did not regain its 1929 peak until November 1954, a quarter century later.

The crash was not the catastrophe. The crash was the first 48 percent. The catastrophe was the second act, the one that took the market down a further 86 percent from the rally high while the dip-buyer averaged down into a monetary vacuum.

Any report titled the way this one is titled must explain why the 2028 to 2029 buyer is the 1932 buyer and not the 1930 buyer. That is the entire question, and it has a precise answer.

Part II: The Autopsy. It Was the Brake, Not the Bubble

The 1930 to 1932 collapse was not the delayed punishment for 1920s speculation. It was the product of policy applied after the crash: a monetary system that was contractionary by construction, operated by a central bank that used its discretion in exactly the wrong direction at every fork.

On the way up, the Fed sterilized. Through the 1920s the United States absorbed the world's monetary gold and neutralized it, refusing to let the inflow expand the money supply, exporting deflation to the periphery and hollowing out the system that would later need to absorb the shock.

On the way down, the Fed tightened. When Britain abandoned gold on September 21, 1931 and the US gold stock fell $727 million, or 15 percent, in six weeks, the New York Fed responded to the outflow by raising its discount rate from 1.5 percent to 3.5 percent in a single week that October: "the sharpest rise within so brief a period in the whole history of the System," delivered into the second year of a depression. Peter Temin calls it one of the most memorable acts of misguided monetary policy in history. In the five months surrounding that decision, 1,860 banks holding $1.45 billion of deposits suspended. From August 1929 to March 1933, the money stock contracted by more than a third. The Fed defended the gold parity and sacrificed the banking system.

Read those two policies as one machine and the 1930 dip-buyer's error becomes visible. He bought an asset class whose monetary foundation was being demolished beneath him: a shrinking money supply, a central bank statutorily and ideologically committed to tightening into outflows, a Treasury running surpluses, and ten thousand banks failing between him and the recovery. His mistake was not timing. His mistake was fighting a system whose brake was welded to the floor.

Now note who prospered. The buyer of July 1932, the one who bought after the brake had done its work and just before Roosevelt took the monetary system off gold, caught the most explosive bull market in American history. The Dow ran from 41.22 to 194.40 by March 10, 1937, a gain of 372 percent in under five years, including the single best year the index has ever printed, 66.7 percent in 1933, the year the brake was finally released.

The lesson of the 1930s is not "never buy crashes." The lesson is that the crash is buyable the moment the monetary regime turns expansionary, and not one day before.

In 1932 and 1933 that turn required a new president, a bank holiday, and the abandonment of gold. The Accord's significance is that the turn has been executed in advance, legislated and staffed and already running, two years before the break this framework expects.

Part III: The Break Still Comes. We Are Not Crash Deniers

Nothing in this series argues the 2020s escape the reckoning. The clock's endgame sequence, restated from the codicil with the amplitudes of the narrative work, runs as follows.

The melt-up runs hotter than 1927 to 1929. The 1920s boom ran against partial resistance, with the Fed sterilizing most of the inflow. The 2020s boom runs against none: an unsterilizable tokenized-dollar inflow, a statutory bill bid, and a Fed whose cuts transmit fully while its hikes feed the machine. Our framework convention doubles the analog amplitudes. Where the Dow returned 28.75 percent in 1927 and 48.22 percent in 1928, the 2026 and 2027 analog years run at roughly twice those rates, carrying the S&P 500 toward roughly 2.5 times its August 2026 level, on the order of 19,000 to 19,400, by the September 2028 peak window.

The break is foreign-sourced. The float is fueled by deposit flight: Chinese deposits of roughly $48 trillion still growing 8 percent a year, Japan's $8.7 trillion, and the 75 to 77 trillion yuan of Chinese time deposits maturing into repression-level rates. That is the same flight that eventually seizes the periphery's banking systems. On the clock, the China and Japan banking strain maps to late 2028, the analog of September 1931. The float stalls or reverses, which is the exit signal, and the US market, the last and highest to fall, breaks on the order of 50 percent.

The window is compressed. The 1929 to 1932 decline took thirty-four months because the brake kept grinding, each false bottom met with sterilization, rate defense, and bank failures. Strip the brake out and the decline is a repricing event rather than a monetary destruction event. Our framework's convention is a peak in the September 2028 window and a terminal low in roughly March 2029: five to six months, the shape of 1929's own first leg, September 3 to November 13, ten weeks, rather than the shape of 1929 to 1932. Everything else in this cycle has run compressed against its analog. The regulatory wave arrived a decade early and the political backlash is being amortized in real time. We expect the bear market to compress the same way.

A 50 percent decline in five months will not feel like a buying opportunity while it is happening. It will feel like 1929, because that is what it rhymes with. The remaining parts explain why it resolves like 1932 and 1933 instead of 1930 to 1932, and why the buyer should act on the schedule of months rather than years.

Part IV: The Floor. Five Reasons the Bottom Holds

One. The flight capital cannot be sterilized, and this time nobody will try. In 1931, foreign crisis meant gold draining out of everyone's banking system into hoards, with each central bank tightening to defend its share. In 2028 and 2029, foreign banking crisis means deposit flight into tokenized dollars, and every tokenized dollar is a statutory buyer of Treasury bills within 93 days. The BIS has already measured the channel's immunity to capital controls: FX restrictions cut traditional deposit dollarization by 25 to 32 percentage points but have no statistically significant effect on stablecoin flows. The worse the periphery's crisis, the stronger the bid under American paper and the easier American financial conditions become. Foreign fragility is mechanically an American subsidy. That was the codicil's line, and it is now load-bearing at the bottom of a crash. The cushion the 1930 to 1931 Fed refused to allow is the 2029 system's default setting.

Two. The Fed has only an accelerator, and the accelerator works. Conventional hikes leak through the stablecoin channel, but cuts transmit fully, and quantities transmit at full old force. Warsh's balance-sheet task force and proposed Treasury accord are institutional machinery for exactly one maneuver: coordinated quantity expansion. A Fed rebuilt around QE, facing a crash, with a chairman on record that AI is "structurally disinflationary" and supplies room to cut, is a Fed that cuts hard and expands fast. The 1930 to 1931 Fed had a doctrine that forbade rescue. The 2028 to 2029 Fed has a doctrine that pre-justifies it.

Three. There are no golden handcuffs, and the dollar is the refuge rather than the casualty. The 1931 Fed hiked into the crash because it had to. Defending the gold parity was the legal and ideological prime directive, and the gold was leaving. The 2028 Fed defends nothing external. There is no parity, no convertibility, no reserve ratio to protect. And in the projected crisis the dollar is not under attack. It is the destination, the unit into which the world's deposits are fleeing. The single mechanism that forced contractionary policy into the teeth of every gold-standard depression has no 2020s counterpart. The Fed of 1931 was handcuffed to the thing draining away. The Fed of 2029 is handcuffed to nothing, and the drain runs toward it.

Four. Fiscal capacity expands in the bust instead of collapsing. The 1930 to 1932 Treasury raised taxes into the depression to defend its credit. The 2028 to 2029 Treasury operates the Genius Move: a marginal funding stack of 13-week bills that reprices to the Fed's cuts within weeks. Every emergency cut lowers the government's interest bill almost immediately, freeing fiscal room at precisely the moment it is needed, the opposite of a long-funded sovereign watching coupons compound through a crisis. The crash loosens the government's constraint rather than tightening it. Counter-cyclical response is pre-funded by the structure of the debt itself.

Five. The vigilante's instrument has been retired. The classical bear-market doom loop of crash, deficits, bond-market revolt, forced austerity, and deeper crash ran through the long end of the curve. That instrument is being decommissioned in plain sight: the 30-year auction frozen at $25 billion since May 2024, long-end buybacks doubled to at least $4 billion per operation and four operations per quarter as of August 19, and roughly 61 percent of surveyed dealers expecting outright reductions in 30-year issuance. By 2029 the marginal dollar of federal debt lives at the front of the curve, priced by policy, bought by statute. There is no maturity left through which the market can impose 1931 on the government. The buyers' strike, the one outcome that turns a crash into a regime change, has been structurally eliminated.

Sum the five and the asymmetry is stated simply. 1930's dip-buyer fought a system designed to contract. 2029's dip-buyer stands with a system that is expansionary by construction, and that expands harder the worse the crisis gets.

Part V: The Arithmetic of the Opportunity

Run the framework's own numbers, stated as conventions rather than forecasts.

The entry. The S&P 500 sits near 7,750 in August 2026. The framework's melt-up convention, roughly 2.5 times by the September 2028 window, puts the peak on the order of 19,400. A 50 percent decline from that peak bottoms near 9,700. Mark the implication, because it is the quiet heart of this report: the projected bottom of the worst crash of the decade sits roughly 25 percent above today's price. On the framework's own arithmetic, the investor who buys the terminal low of the 2028 to 2029 collapse pays more than the investor who simply holds from August 2026. The melt-up is that large. The crash is not the risk to fear. Missing the sequence is.

The precedents for buying the released brake. Two entries in the last century combined crash prices with an expansionary regime, and both were generational. The 1932 buyer, entering after the brake had done its work, caught 372 percent in under five years, with 66.7 percent in 1933 alone as the gold exit released the spring. The March 2009 buyer, entering into a Fed running QE at full throttle, earned a 17.8 percent annualized total return over the following decade and roughly six times his money over fifteen years. The 2029 entry is, on this framework, both at once: 1932's washed-out prices with 2009's policy stance, plus a structural, statutory, price-insensitive bid that neither predecessor enjoyed and that grows through the crisis as periphery flight capital converts to float.

The recovery shape. The framework expects the rebound to begin within the 2029 window and hold, a V where 1929 to 1932 carved an L, with recovery to the vicinity of the prior peak by roughly December 2029 and the subsequent expansion running into the 2030s on the tokenized-dollar standard. The dollar emerges as the sole standing monetary standard, what the codicil called the 1944 outcome without the conference, and the assets denominated in it reprice for that world. That is why the title says the buying opportunity of the next decade: not because a 50 percent discount is rare, but because a 50 percent discount inside a monetary regime that can only expand is, on the historical record, the single best configuration equities ever offer.

What to buy. The allocation logic of the prior reports survives the crash and strengthens at the bottom. Own the users of AI and stablecoin productivity, meaning the Dow Industrials, the Transports, the Russell 2000, and the equal-weight and real-economy S&P, over the producers of compute, for the same reason the theater chains outran the set-makers after 1927. At the low, add the direct beneficiaries of the recovery mechanics themselves: the domestic real-economy franchises that a bill-funded, QE-cushioned, flight-capital-financed expansion reflates first. The compute complex will bottom too, but it carries the capex and counterparty scars of the bust. Oracle down 32 percent and CoreWeave down 38 percent in the 2026 drawdowns were the preview. The users carry the productivity gains without the write-offs.

Part VI: The Checklist. Telling the Buy from the Trap in Real Time

A framework that cannot specify its own failure is a faith. Here is how the 2028 to 2029 buyer distinguishes 1932 from 1930 while the screen is red, in order of importance.

First, the float. The stablecoin float is simultaneously the fuel gauge and the exit gauge. If the float grows through the crash, with periphery flight capital converting to tokenized dollars faster than domestic holders redeem, the cushion is live and the floor thesis is intact. If the float contracts through the crash, the statutory bid is liquidating instead of buying, the cheap leg of the sovereign trade is failing, and this framework is wrong. Stand aside, because that world is 1930. This is the single decisive variable. Watch it weekly.

Second, the Fed's first move. The floor thesis requires cuts and balance-sheet expansion within weeks of the break, consistent with the machinery Warsh has built. Any 1931-style gesture, meaning tightening or passivity in the name of discipline while the periphery burns, is disconfirming. Warsh's pre-commitment against stablecoin bailouts, "We do not want to be in the bailout business, full stop," is the stated stress point. The framework expects that stance to be tested against a run that the deposit-base arithmetic makes unlikely, not impossible.

Third, the bill auctions. The Genius Move's crisis performance is measured at the front of the curve: bills covered, yields pinned near policy, the Treasury's funding machine untroubled. Failing bill auctions during the crash would mean the duration-in-the-float alchemy has broken, and that is disconfirming.

Fourth, the geography of the crisis. The floor thesis requires the banking crisis to stay foreign, with China and Japan seizing, the dollar strengthening, and flight capital inbound. An American deposit crisis, or a dollar falling during the panic, inverts the mechanics and voids the analog.

Fifth, the entry schedule. Conditional on the first four confirming, scale in across the window rather than calling the low: tranches from the first 35 percent drawdown print through the March 2029 convention, completing by the time the Fed's quantity response is visibly in the tape. The 1932 lesson cuts both ways. The reward went to those who bought when the regime turned, and the regime turn, this time, is scheduled to be simultaneous with the crash rather than three years behind it.

Coda: The Honest Ledger

For the record, what this report does not claim. The Bessent/Warsh Accord is our name for an observable policy convergence, not a signed agreement. The 99-year clock is an analytic convention, and September 2028 and March 2029 are its coordinates, not appointments the market has agreed to keep. The plumbing beneath the floor thesis has never been tested at scale: a statutory bill bid absorbing a genuine float reversal, a no-bailout chairman facing a live run, a bill-funded sovereign rolling $7 trillion through a panic. The first test will be run live. The consensus disagrees with the melt-up amplitudes, the compression, and the V. If the float stalls before the peak window, the entire sequence of melt-up, break, and floor is voided, and the prior reports' exit discipline applies.

But the core of the matter survives every caveat. The difference between 1930 and 1932 was never the price of stocks. It was the direction of the machine behind them. In 1930 the machine was contracting and the dip-buyer was crushed. In 1932 the machine turned expansionary and the buyer of the same wreckage made 372 percent. The 2020s have done something without precedent: they turned the machine expansionary in advance, by statute, and then removed the lever by which it could be turned back. When the break comes, and the clock says late 2028, the investor will be offered the prices of 1932 with the policy of 2009, financed by the flight capital of a world whose money has nowhere else to go.

That is the buying opportunity of the next decade. The work between now and then is watching the float.

JD Unfiltered. Joseph M. Salvani and Daniel J. Walsh.

A JD Unfiltered Market Analog Report, Part III of the Bessent/Warsh Accord series, extending "The Second Twenties" (August 10, 2026), "The Bessent/Warsh Accord" codicil (August 10, 2026), and "Genius Move" (August 20, 2026). This document presents a historical-analog market thesis for discussion purposes and is not investment advice. The "Bessent/Warsh Accord" is the author's name for a described policy convergence, not a claimed formal agreement. All projected dates, magnitudes, and price levels are illustrative conventions of the analog framework, not forecasts. Historical parallels are illustrative frameworks, not predictions.

The authors hold positions in securities mentioned and reserve the right to buy or sell shares at any time without notice.