Japan Currency Outlook: The End of Extreme Yen Weakness

On Thursday 30 July the dollar reached ¥163.73, the yen’s weakest level in approximately four decades. By Friday’s close it stood at ¥157.57, the dollar’s largest single day decline since 2022. On Monday 3 August, Finance Minister Satsuki Katayama confirmed that Japan’s Ministry of Finance had conducted a yen buying operation on Friday 31 July jointly with the US Treasury, executed pursuant to the US Japan Finance Ministers’ Joint Statement of September 2025 and directed at excessive volatility and disorderly movements in the yen. Secretary Bessent confirmed the operation in parallel terms, and both authorities stated they would not hesitate to intervene jointly again.

The significant detail is not the level. It is the joint execution. This was the first coordinated US and Japanese purchase of yen since 1998, an interval of twenty eight years. The 2011 G7 action, frequently miscited as the more recent precedent, ran in the opposite direction: it sold yen in order to weaken a currency appreciating on post tsunami repatriation flows.

Price action through the subsequent week requires acknowledgement. The yen surrendered close to half of its intervention driven gains by Thursday 6 August, weakening back through ¥158 and prompting renewed speculation regarding further official action. This is the standard post intervention pattern and we do not regard it as inconsistent with our thesis. Friday’s US employment data then reversed the move, and the pair closed the week around ¥157.7 to ¥158.5. That employment report is material to the outlook and we address it in Section VII.

Between 28 April and 27 May 2026 the Ministry of Finance intervened unilaterally in size. Within six weeks USD/JPY had traded back through the level at which the operation commenced.

The explanation is mechanical. Direction determines the funding constraint. When Japan sells yen to weaken the currency, the Ministry of Finance issues financing bills and the yen leg is effectively unconstrained. When Japan buys yen, it must liquidate foreign currency assets held in the Foreign Exchange Fund Special Account. Those assets are predominantly US Treasury securities.

The reserve data record the cost, though they require careful reading. Japan’s official reserve assets stood at USD 1,374,734m at the end of March, rose to USD 1,382,994m at the end of April, fell to USD 1,305,874m at the end of May and declined further to USD 1,287,476m at the end of June. That is a cumulative change of approximately USD 95bn across the two months following the intervention.

We would caution against interpreting that figure as intervention spending in full. Reserve assets are marked to market and non dollar holdings are translated at prevailing rates, so monthly movements combine operations with valuation effects. The March decline of USD 35,965m, for instance, occurred with no intervention reported at all. The cleaner evidence lies in the composition: Bloomberg reported that Japan’s foreign securities holdings declined USD 75.6bn in May alone, which gives an indication of the scale of the operation. US data show Japan holding approximately USD 1.1trn of US Treasuries as of May 2026, the largest foreign holding.

Press estimates put the 31 July operation in the range of USD 60bn to 80bn. The Ministry has not yet published a figure. July reserve data is due on the seventh business day of August and will provide the first official read, with the Ministry’s periodic intervention disclosure to follow.

Each unilateral defence therefore carried an embedded self defeating mechanism. Tokyo sold Treasuries, US yields rose, the differential driving the carry trade widened, and a portion of the pressure Japan sought to relieve was regenerated by the act of relieving it.

In our view this is why Washington participated, rather than alliance considerations. The US Treasury has a direct interest in Japan not becoming a persistent forced seller of duration into a market absorbing record issuance. Purchasing yen alongside Tokyo removes the need for Tokyo to raise dollars through bond sales.

The Ministry of Finance statement of 3 August confirms this reading. Alongside the intervention, Japan announced it plans to use the Federal Reserve’s Foreign and International Monetary Authorities repo facility, which allows an approved monetary authority to borrow dollars against Treasury collateral rather than selling it, subject to a limit of USD 60bn per counterparty per day. Secretary Bessent has since publicly urged the Federal Reserve to expand the facility, stating that he wants it upsized. Any expansion requires approval from the Federal Open Market Committee and none has been announced.

Taken together, the joint operation and the FIMA arrangement constitute a Treasury market financing mechanism executed in foreign exchange form. State Street’s Masahiko Loo has argued that the FIMA signal may matter more than the intervention itself. We agree.

The constraint that caused every prior defence to fail has been lifted.

The joint operation did not terminate the carry trade. It arrived at a point when the trade was already failing, which is why it proved effective.

Yen denominated foreign currency credit to non residents is the closest available proxy for the funding leg of the global carry trade. The BIS global liquidity indicators record the following sequence of year on year growth rates: 6% at the end of the first quarter of 2025, zero at the end of the second, a contraction of 4% at the end of the third, and a contraction of 4.9% across calendar 2025 as a whole. The BIS attributes the slowdown, which began after the third quarter of 2024, to the onset of Japanese monetary tightening and the August 2024 unwinding of carry positions. Growth in the preceding two years had run well above 10%.

The comparison with the other two major funding currencies is the point. Over the same five quarters, dollar credit to non residents grew 5%, 6%, 7%, 8.5% and most recently 7.3%, reaching USD 14.7trn. Euro credit grew 10%, 13%, 11%, 11% and most recently 12%, reaching EUR 5.1trn. The yen alone has gone from expansion to contraction while its peers accelerated.

The BIS published its end March 2026 indicators on 31 July, the day of the intervention. That release reports the dollar and euro growth rates cited above. It states no figure for the yen. When a statistical release ceases to report a series it has consistently published alongside its peers, the omission itself is informative, and we would treat the absence as directionally consistent with the preceding four quarters rather than as an improvement.

Against a stock of borrowed yen that has been shrinking rather than growing, speculative futures positioning stood at 152,125 contracts net short as of 21 July, approximately USD 11.7bn notional and an extreme by the standards of the past two years. A crowded short constructed on a contracting funding base is structurally unstable. This is the condition under which official flow achieves asymmetric leverage, and the Ministry of Finance selected its timing accordingly.

The consensus explanation for USD/JPY remains the interest rate differential. That explanation is failing on its own terms. Lazard Asset Management notes that the differential accounted for roughly 90% of the variance in USD/JPY through 2023, and considerably less subsequently. Through 2026 the yen weakened while the Bank of Japan tightened. The policy rate reached 1.00% in June, a level not seen since 1995, and the currency proceeded to a four decade low regardless. A variable that moves contrary to its own explanation is not the explanatory variable.

Richard A. Werner has argued since the early 1990s that the operative quantity is credit creation rather than its price, and that interest rates follow nominal activity rather than lead it. Applied to currency pairs, the relevant measure is the relative credit creation of the two jurisdictions. We would emphasise that the central bank leg carries this argument. The commercial bank data is corroborative rather than decisive, and we set out its limits below.

Central bank credit creation. The Federal Reserve concluded balance sheet runoff in November 2025 and commenced reserve management purchases in December, initially approximately USD 40bn of Treasury bills monthly, with agency MBS runoff additionally reinvested into bills. Whatever terminology has been avoided, the Federal Reserve balance sheet is expanding. The Bank of Japan’s is not. Werner’s Leading Liquidity Index, a proprietary series that third parties cannot independently reconstruct, shows Bank of Japan credit creation substantially negative while the Federal Reserve’s has turned positive. We cite it as his published measure rather than as verified data. The direction of the two balance sheets, however, is a matter of public record, and it points toward a stronger yen and a weaker dollar.

Commercial bank credit creation. Japanese bank lending accelerated through the first half of 2026. Growth ran at 4.4% year on year in December 2025 and 4.5% in January and February, rose to 4.8% in March with outstanding loans at ¥667.0trn, and reached 5.7% in both May and June, the fastest expansion since March 2021, with loans at ¥670.8trn and ¥676.1trn respectively. The June figure excluding shinkin banks was 6.3%. The Bank of Japan attributes the expansion to merger and acquisition activity, real estate lending and broader economic recovery.

The July release, published on 7 August, showed growth easing to 5.4%, with outstanding loans at ¥679.2trn and the major and regional bank measure slowing to 5.9% from 6.3%. We regard a single month as insufficient to establish a turn, but note it as the first deceleration of the sequence. We would also note that published sources conflict on the April reading, reporting either 5.4% or 4.8%, and we therefore exclude that month rather than choose between them.

The significance is the level shift rather than any comparison with the United States. Japan is emerging from two decades in which credit growth hovered near zero and is now sustaining rates above 5%. On the American side, loans and leases at all commercial banks grew 5.0% in October 2025, 5.4% in November, 5.6% in December, 6.2% in January and 6.7% in February 2026, which is the most recent monthly observation we have been able to verify. We therefore make no claim about the relative position of the two series in the second quarter. The Japanese data stands on its own: this is genuine real economy recovery, which is also the reason Japanese interest rates were required to rise, rather than the reverse.

Werner’s own empirical work across the US, Japan, the UK and Germany finds interest rates following nominal GDP growth positively and in the same direction, which if correct implies that Bank of Japan gradualism is a lagging accommodation to activity rather than a driver of it.

Conventional analysis holds that currency weakness assists an export led economy. For an economy that has already ascended the value added ladder, we regard the opposite as correct.

An advanced economy imports low value added inputs and exports high value added output. Demand for genuinely differentiated goods is relatively price inelastic, so devaluation generates limited incremental export volume while rendering necessary imports unambiguously more expensive. Terms of trade, defined as export prices divided by import prices, deteriorate.

The second order effect is behavioural and more damaging. Currency weakness delivers translated foreign exchange profits absent operational improvement. It relieves precisely the pressure to innovate, re tool and reinvest that established the value added position initially. Margin arrives without effort, and the effort ceases. Shareholders extract the windfall as dividends and capital expenditure is deferred.

Japan and Germany prospered across decades under appreciating currencies. The appreciation constituted the discipline. Germany’s experience following the abolition of the Deutsche Mark is, in our reading, consistent with this argument, though we would note that German underperformance is overdetermined by energy costs, China exposure and demographics, and we do not claim currency weakness as the sole cause.

Four decades of yen weakness therefore represent accumulated damage to Japanese corporate competitiveness rather than accumulated advantage. Tokyo’s persistence in defending the currency is economically rational.

The single most consequential policy development of the summer received a fraction of the coverage devoted to the intervention itself, and in our assessment it will prove more durable.

On 10 July 2026, Finance Minister Satsuki Katayama stated that the government wished to pursue measures encouraging pension funds, including the Government Pension Investment Fund, to make substantially greater investments in Japanese financial assets. The GPIF managed ¥293.6trn, approximately USD 1.8trn, at the end of March, and its basic portfolio currently allocates 25% each to domestic bonds, foreign bonds, domestic equities and foreign equities. The market response was immediate: the yen appreciated approximately 0.6% to 161.285, and the ten year JGB yield fell 11.5 basis points to 2.760%, its steepest single day decline in more than a year.

The context is that this reverses a deliberate policy. In 2014, facing an ageing population and negligible domestic returns, the Abe administration directed the GPIF away from its conservative domestic bond allocation and toward overseas assets. That decision initiated more than a decade of Japanese capital outflow and constituted a significant component of the structural bid for foreign, principally US, securities. Katayama’s remarks signal the intention to reverse it. HSBC has characterised asset repatriation as the missing piece in Japan’s reflation programme.

Two observations follow, and we regard the second as the more important.

First, the arithmetic threshold has been crossed. Domestic institutions historically purchased foreign bonds because JGBs offered no yield. The level at which domestic bonds were widely considered to become competitive again was a ten year yield in the region of 1.75% to 1.77%. The ten year opened 2026 just below 2.1% and exceeded 2.8% in May, the highest since 1996, before settling near 2.7%. The hurdle rate for overseas allocation has risen materially, and the incentive for repatriation now exists independently of any government encouragement.

Second, and more revealing, is the channel Tokyo has chosen. The Takaichi administration has consistently resisted Bank of Japan tightening. The Prime Minister opposed rate increases before taking office and, according to reporting in February, conveyed concerns regarding further hikes directly to Governor Ueda. Yet the same administration is actively pursuing capital repatriation and a firmer currency through the allocation of institutional balance sheets.

Tokyo is operating on quantities while declining to operate on prices. The authority closest to the problem is not behaving as though the interest rate is the instrument that matters.

The policy differential remains approximately 275 basis points, with the Federal Reserve at 3.50% to 3.75% and the Bank of Japan at 1.00%. The July FOMC held for a fifth consecutive meeting on a 9 to 3 vote, with three regional presidents dissenting in favour of a hike, the first unidirectional triple dissent since September 2016. The carry trade has re established itself following each intervention of the past four years, and the partial retracement of the past week is consistent with that history.

The July employment report published on Friday 7 August materially alters this risk profile. US nonfarm payrolls declined by 23,000 against consensus expectations of a gain in the region of 83,000 to 95,000, the first outright monthly contraction of the cycle. Revisions removed a combined 103,000 jobs from the May and June readings, with May reduced from 129,000 to 63,000. The three month average is now approximately 20,000. Average hourly earnings growth fell to 3.2% year on year, the lowest since May 2021. The decline in the unemployment rate to 4.1% reflects labour force exit rather than hiring. We would note that private payrolls rose 30,000 and the headline was driven by a 53,000 decline in government employment, so the print is less severe than it appears; nevertheless the trend is unambiguous.

The principal risk to our thesis was a September Federal Reserve hike. That outcome is now considerably less probable.

We would revise our view on any of the following:

• BIS second quarter 2026 yen credit returning to positive growth.
• A September Federal Reserve hike notwithstanding the labour data.
• CFTC net short positioning rebuilding beyond 150,000 contracts.
• Japanese bank lending decelerating below 5%, the July reading of 5.4% having already moved in that direction.
• The GPIF allocation review concluding without material change to the foreign asset weighting.

Base case, sustained reversal (55%). Yen appreciation toward ¥140 to ¥145 over twelve months, with ¥130 to ¥140 achievable should the credit differential widen further. This assumes Bank of Japan hikes in September and again by the first quarter of 2027. Consensus terminal rate estimates have moved to 1.75%, with certain houses at 2.5%. Japanese corporates operated effectively at ¥80 within recent memory; ¥140 does not constitute a strong yen by any historical standard.

Range consolidation (30%). ¥150 to ¥160 establishes as a new band, with repeated joint operations capping the upper bound. This represents a slower path to the same destination rather than a refutation of the thesis, and the price action of the past week is consistent with it.

Resumed weakness (15%). A return through ¥163 within six months, requiring a Federal Reserve hike alongside Bank of Japan hesitation. We regard this as materially less probable following Friday’s labour data.

Portfolio implications. Unhedged foreign holders of Japanese equities have been collecting a currency tailwind that now reverses; the hedging decision carries greater consequence than the allocation decision from this point. Japanese domestic financial institutions represent the cleanest expression of a credit cycle turning after twenty years, and would additionally benefit from any repatriation of institutional assets into domestic markets. Holders of long dated US Treasuries should note that the repatriation channel and Japan’s reserve management operate on the same side of that market, and that the FIMA arrangement exists precisely to blunt it. Any strategy carrying a short yen funding leg should be sized on the assumption that the July squeeze is the first of several rather than the last.

The most instructive fact of the past twelve months is one that conventional analysis cannot accommodate. The Bank of Japan raised its policy rate to a level unseen since 1995, and the yen proceeded to a forty year low. Under the textbook account of exchange rate determination this sequence should not occur. It occurred anyway, and it occurred over a sustained period rather than as a transient anomaly.

The error is one of category. Interest rates are treated as the mechanism when they are more accurately a symptom, reflecting nominal activity rather than determining it, and following that activity positively rather than inversely. Analysis built on the price of credit will therefore misread turning points systematically, because it is observing an output of the system and modelling it as an input. A framework that could not explain why a tightening central bank presided over a collapsing currency is unlikely to explain why that currency now recovers.

The variable that does the explanatory work is the quantity of credit created, and specifically its distribution between the two jurisdictions in a currency pair. On that measure the evidence for a turning point is substantial: a Federal Reserve balance sheet that is expanding again against a Bank of Japan balance sheet that is not, Japanese bank lending sustained above 5% after two decades near zero, and the yen denominated funding stock supporting the carry trade contracting in every quarter for which the BIS has published a figure since the second quarter of 2025. This proposition has been available in the empirical literature for more than three decades. It has been disregarded for most of them, at considerable cost to those who disregarded it.

What changed on 31 July is that the constraint binding Japanese intervention, namely the requirement to sell Treasuries in order to buy yen, was removed by Washington for reasons rooted in the US bond market. What changed on 10 July, with less attention, is that Tokyo began assembling the institutional flows to make a firmer currency self sustaining.

We regard the era of extreme yen weakness as ending. It was a policy rather than an accident, and it has outlived its usefulness to those who imposed it.