Japan does not lose money defending the yen. It books a spread. When the Ministry of Finance sold roughly $95 billion of foreign reserves in August, the...
Japan does not lose money defending the yen. It books a spread. When the Ministry of Finance sold roughly $95 billion of foreign reserves in August, the reflexive read was distress selling: a sovereign burning through its dollar hoard to slow a currency it cannot control. The more accurate read is that Tokyo was closing a position it had financed for a decade at near-zero cost and is now unwinding into high dollar yields. The $31 billion of intervention profit the government is already quarreling over from last fiscal year is not an accident of timing. It is the structural payoff of the trade.
The Carry Embedded in the Reserve Stack
Foreign exchange intervention that sells dollars for yen has an unusual property for a central operation: it is long the very yield differential that is driving the currency it is trying to defend. Japan's reserves are overwhelmingly parked in US Treasuries. Those Treasuries were accumulated across years when the funding side, yen liabilities issued by the MOF through financing bills, cost almost nothing. The 10Y is at 4.78% and the 2Y at 4.37% as of early September. The asset leg throws off coupon income at those levels; the liability leg was struck when Japanese short rates sat at or below zero.
That is a positive carry position of enormous size, and intervention is the mechanism that realizes it. Every time the MOF sells a slug of Treasuries to buy yen, it crystallizes the accrued gain on paper bought cheaper and yielding more, plus the currency move if the yen has weakened since purchase. The $95 billion August drawdown is the visible flow. The $31 billion profit is the residue. Both are the same trade viewed from opposite ends.
The consensus framing misses this because it imports a corporate-treasury intuition, where selling your reserves signals stress. A sovereign that issues the funding currency and holds the high-yield asset is in the opposite posture. It is the counterparty everyone else pays.
Why the Second-Order Move Matters More Than the Headline
The policy headline is "Japan intervenes to support the yen." The transmission that matters is what the Treasury leg does to the marginal dollar-rate bid.
Japan is one of the largest foreign holders of US government debt. When the MOF liquidates $95 billion in a month, that supply does not vanish into a vacuum; it lands on a Treasury market already absorbing heavy issuance at a 4.78% 10Y. Intervention that props the yen is simultaneously a small tightening impulse on the long end of the US curve. The two legs are not independent. The same act that strengthens the yen nudges US term premium in the direction that widens the differential intervention is meant to lean against.
This is the reflexive knot at the center of the trade. A currency defense funded by selling the high-yield asset marginally raises that asset's yield, which marginally supports the case for the weak currency it was defending. It does not unwind the position; it makes the carry richer at the next intervention. The profit compounds precisely because the intervention is partly self-defeating on the exchange rate while being self-reinforcing on the spread.
The 2s10s at +0.41% tells you the curve is no longer inverted, which changes the reinvestment calculus. Reserves rolling off can be redeployed at a positive term slope rather than into an inverted curve where the long end paid less than cash. The reinvestment yield on the reserve stack is improving even as the stack shrinks.
The Fight Over the Profit Is the Real Signal
Follow the money quarrel, not the intervention notice. A government does not argue over how to spend $31 billion it expects to give back. The internal fight over allocating the intervention gain is a tacit admission that Tokyo treats these operations as revenue, not as an emergency drawdown to be replenished at a loss.
That reframes the whole exercise. If intervention were genuinely a defensive expenditure, the accounting question would be how to fund the shortfall. Instead it is how to distribute the surplus. The behavior reveals the belief: the MOF expects the carry to keep paying, which means it expects the rate differential, and therefore the pressure on the yen, to persist. The profit and the weak-yen problem are the same phenomenon.
For a cross-asset reader, that is the actionable inference. The party best positioned to know the durability of the dollar-yen differential is booking it as recurring income and dividing the proceeds. Revealed preference beats forward guidance.
What Would Break This Read
The thesis is not immune. It rests on the funding leg staying cheap and the asset leg staying rich, and both can move.
The clean break is a Bank of Japan hiking cycle that lifts the yen funding cost toward the dollar asset yield. Compress the differential and the carry that makes intervention profitable erodes from the liability side. At that point selling Treasuries to buy yen stops being a spread harvest and starts being a genuine cost, because the reserves are no longer funded at zero. A sustained rise in Japanese short rates, not a single symbolic hike, is the condition that inverts the economics.
The second break is on the asset leg: a sharp US rally that drives the 10Y well below the level at which the bulk of the reserve stack was accumulated. That would turn crystallized gains into crystallized losses on the marginal sale. With the 10Y at 4.78%, most of the stack is comfortably in the money, so this is the less immediate risk, but it is the one that would flip the headline from "profitable intervention" to "reserve loss" fastest if a US recession forced aggressive cuts.
There is also a fair version of the bearish case that does not require rates to move at all: scale. Reserves are finite. A $95 billion month is not repeatable indefinitely, and the market knows the ammunition has a floor. But that argues the intervention's currency effect fades, not that it becomes unprofitable. Running out of a winning trade is a different problem from running a losing one.
Where Confirmation Shows Up Next
Watch the spread between Japanese short rates and US Treasury yields, because that spread is the profit-and-loss statement of the entire operation. As long as it stays wide, every intervention is a spread harvest dressed as a defense, and the internal fights over how to spend the proceeds will recur.
The read inverts the moment the BoJ signals a funding cost that closes the gap. Until Japanese short rates rise enough to make the yen leg expensive, don't read reserve drawdowns as distress. Read them as Tokyo selling a rich asset it financed for free, and counting the difference.
