Japan remains the largest foreign holder of U.S. Treasury securities, with holdings still above $1.1 trillion. For decades, that position helped finance American deficits at relatively contained rates. The arrangement, however, is fraying.
Read more What the Flock! How One Company And Its ALPRs Enable Warrantless Surveillance
The Bank of Japan has and is gradually stepping back from the extreme yield-curve control that kept Japanese government bond yields artificially suppressed for years. Domestic Japanese yields have risen. The yen has been under pressure, prompting repeated intervention. When Tokyo defends the yen, it often draws on dollar reserves that include Treasuries. The net result is reduced Japanese demand, or outright sales, for U.S. government debt at the precise moment the United States must roll over enormous volumes of maturing obligations against a debt stock that has crossed $39 trillion.
This is a structural shift. Foreign official buyers are no longer the reliable residual demand they once were. Domestic private buyers and the banking system will have to absorb more paper. Markets have already tested higher long-term yields. The political system will not tolerate market-clearing rates high enough to clear the entire supply on purely commercial terms for long.
The available middle road has a name and historical precedent. “Financial repression” is what governments do when default is unacceptable and honest fiscal consolidation is politically impossible: rate caps, regulatory pressure on institutional buyers, and tolerance for above-target inflation, used together to hold the real return on debt below what an open market would demand.
History supplies the template. From 1942 to 1951, the Federal Reserve and Treasury maintained an explicit peg on long-term government yields to finance wartime and early postwar debt, deliberately compressing real returns to bondholders so the government could service and gradually inflate away a large debt burden. The 1951 Treasury-Fed Accord ended the formal arrangement, but the episode demonstrated that when sovereign debt becomes politically non-negotiable, authorities prefer to manage the price of money rather than accept the full market verdict on the debt.
Something similar is the most probable path over the next several years. Near-term auctions may clear at higher yields. Persistent pressure will then produce a mix of balance sheet support, regulatory nudges that increase structural demand for Treasuries, and greater tolerance for inflation that erodes the real value of the debt. Pure default remains extremely unlikely. Pure fiscal consolidation on a scale sufficient to restore confidence is equally unlikely under current political incentives. The middle road is repression dressed in modern institutional clothing.
Two modern forces will shape how long that middle road has to be traveled.
First is artificial intelligence, which cuts against itself across two different time horizons. In the near term, A.I. is inflationary. The energy intensity of data-center buildout, competition for chips and skilled labor, and the sheer scale of capital expenditure are already pushing up costs in the sectors A.I. touches most directly. That is the barrier phase, and it is likely to persist for the next several years.
The longer-term case is stronger than the standard “if A.I. delivers productivity gains” framing suggests, because A.I.’s contribution to growth does not wait for diffusion. As A.I. investment, infrastructure, and output become an industry in their own right, chips, data centers, model development, and applied tooling show up directly in GDP before the broader economy absorbs A.I.-driven efficiency gains in existing white-collar work, logistics, and manufacturing. The productivity effect skeptics look for downstream, in measured output per worker across legacy industries, is a second and slower wave. The first wave is simply the scale of the new industry itself, and it is already underway. Higher nominal GDP from either source makes the existing debt stock more sustainable and reduces the intensity of repression required to manage it. A.I. is therefore a near-term source of inflationary stress and a longer-term, and probably larger, source of relief, arriving in two stages rather than resolving into a single verdict.
Read more Los Angeles County DA Hochman Is Doing The Right Thing
Second is cryptocurrency, particularly Bitcoin. Negative or compressed real rates and visible debt monetization increase demand for assets outside the sovereign monetary system. Bitcoin’s fixed supply makes it a direct competitor to financially repressed cash and bonds. Capital that loses confidence in the long-term real return of Treasuries will seek portable, non-sovereign stores of value. That capital flight itself becomes a political feedback loop: The more visible the success of the alternative, the greater the pressure on authorities to respond.
The American response so far has been revealing. A retail central bank digital currency, the pure Fed-issued digital dollar, faces a statutory prohibition through the end of 2030 as well as deep political resistance on privacy and surveillance grounds. Instead, policy has embraced regulated private dollar stablecoins under the GENIUS Act framework. These instruments must be backed one-to-one with high-quality liquid assets, predominantly short-term Treasuries and cash equivalents. The architecture creates structural demand for government debt while keeping issuance in private hands under regulatory oversight. It is a hybrid solution: private-sector form, public-asset substance, and an explicit attempt to keep capital inside the dollar system rather than letting it migrate entirely into permissionless alternatives.
In short, the United States is already building the infrastructure for a milder, more market-compatible form of repression. Stablecoin growth supports the short end of the curve. Banks and pensions can be encouraged, through capital rules and supervisory guidance, to hold more duration. The Fed retains the capacity to expand its balance sheet if market function deteriorates. The goal is not to eliminate higher nominal yields entirely, but to prevent them from rising to levels that force painful fiscal choices or threaten financial stability.
Savers and holders of long-duration fixed-income assets will bear the cost. Real returns will be lower than what a pure market outcome would have produced. Equity and real assets, including scarce digital assets, will tend to outperform relative to cash and bonds during such periods, the classic pattern under financial repression. The classic consequences will be felt by the classic victims.
The Japan dynamic does not guarantee crisis, but it does foreclose the assumption Washington has quietly relied on for a decade: that foreign official demand would remain the reliable buyer of first resort. Once that assumption fails, some form of financial repression follows almost by necessity. The political system will not choose default, and it will not choose the spending discipline required to avoid repression through consolidation alone. What remains open is not whether repression happens, but how long it has to last and how visible it has to be.
A.I.’s two-stage growth effect works to shorten the window in which repression is needed at all. Sustained capital flight into Bitcoin and other assets outside the sovereign system raises the political cost of repression that is too overt, pushing authorities toward the softer, more market-compatible version already visible in the Stablecoin architecture.
Together, the two forces determine repression’s severity and duration. That, more than any single Federal Reserve decision, is what will define the next decade of American monetary policy.
Read more The Navy and the Midterms
Image via PickPik.